7 topics. Each one opens with its key idea, then the detail you need to use it.
1. Verifying solutions
Key idea: Differentiate, substitute into both sides, and simplify; a solution must match for every x, not just one point.
To verify that a function solves a differential equation, compute every derivative the equation needs. Then substitute the function into the left side and the right side separately and compare. Take $y = 4e^{-3x}$ and the equation $\frac{dy}{dx} = -3y$. The left side is $-12e^{-3x}$. The right side is $-3(4e^{-3x}) = -12e^{-3x}$. They match for every $x$, so it is a solution.
Now test $y = x^3$ in $\frac{dy}{dx} = xy$. The left side is $3x^2$, but the right side is $x \cdot x^3 = x^4$. These are different functions, so $x^3$ fails, even though both sides happen to equal 0 at $x = 0$.
The same method works for second-order equations and implicit relations. For $y'' = -9y$, the function $y = \cos(3x)$ gives $y'' = -9\cos(3x)$, which matches. For a relation like $x^2 - y^2 = 7$, differentiate implicitly to get $2x - 2y\frac{dy}{dx} = 0$, so $\frac{dy}{dx} = \frac{x}{y}$. That tells you which equation the relation solves.
Watch the chain rule. The derivative of $e^{x^2}$ is $2xe^{x^2}$, not $e^{x^2}$. Most wrong answers in this topic come from a skipped chain rule or from checking only the shape of a formula instead of substituting it.
2. Slope fields: drawing and reading them
Key idea: Evaluate the slope at chosen points, then use columns, rows, zero-slope lines and signs to match a field to its equation.
A slope field is a grid of short segments. Each segment has the slope that the differential equation gives at that point. To draw one, evaluate $\frac{dy}{dx}$ at each grid point and draw a segment with that slope. A slope of 0 is flat, a slope of 1 rises at 45 degrees on equally scaled axes, and a slope of $-2$ falls steeply.
Reading a slope field means working the other way. Use these checks to match a field to its equation. If the slope depends only on $x$, as in $\frac{dy}{dx} = x^2 - 1$, every segment in the same vertical column is identical. If the slope depends only on $y$, as in $\frac{dy}{dx} = y - 2$, every segment in the same horizontal row is identical. Rows of flat segments then mark equilibrium solutions, here the line $y = 2$.
Next, find where the slope is zero by setting the right side equal to 0. For $\frac{dy}{dx} = x - y$, the flat segments lie along the line $y = x$. Then check signs. Above that line the slope is negative, and below it the slope is positive.
Finally, test points on the axes. On the $y$-axis $x = 0$, and on the $x$-axis $y = 0$. Plugging those in usually eliminates two or three answer choices at once. To sketch a particular solution, start at the given point and follow the segments smoothly. Never cut across them.
3. Separation of variables
Key idea: Separate, integrate with one +C, plug in the initial condition, then solve for y; a constant inside an exponent becomes a multiplier.
Separation of variables works when you can write $\frac{dy}{dx}$ as a function of $x$ times a function of $y$. Move every $y$ term, including $dy$, to one side and every $x$ term to the other. Then integrate both sides. Here is the full routine for $\frac{dy}{dx} = xy^2$ with $y(0) = \frac{1}{2}$.
Step 1, separate: $\frac{dy}{y^2} = x\,dx$.
Step 2, integrate both sides: $-\frac{1}{y} = \frac{x^2}{2} + C$. One constant on one side is enough, because two constants combine into one.
Step 3, use the initial condition now: $-2 = 0 + C$, so $C = -2$.
Step 4, solve for $y$: $-\frac{1}{y} = \frac{x^2}{2} - 2$, which gives $y = \frac{2}{4 - x^2}$.
Never substitute the initial condition before you integrate, because the equation still contains derivatives at that stage. Substituting right after integrating, as in Step 3, is safest because the algebra is easiest there.
When you integrate $\frac{1}{y}$, you get $\ln|y|$. For $\frac{dy}{dx} = ky$, that gives $\ln|y| = kx + C$, so $|y| = e^{C}e^{kx}$. The positive constant $e^C$, together with the plus-or-minus sign from the absolute value, becomes a single constant $A$. That is why the constant multiplies the exponential: $y = Ae^{kx}$, not $e^{kx} + C$. Check your answer by plugging it back into the original equation.
4. General solutions, particular solutions and their domains
Key idea: The domain of a particular solution is the largest open interval around the initial x-value where nothing breaks.
The general solution contains the arbitrary constant, like $y = Ce^{-x} + 2$ for $\frac{dy}{dx} = 2 - y$. Every value of $C$ gives a valid curve. Notice where the $+2$ comes from. Setting $\frac{dy}{dx} = 0$ gives $y = 2$, the equilibrium, and every solution approaches that line. A family with $-2$ in place of $+2$ fails verification, because its curves would level off at the wrong height.
A particular solution has a specific constant, fixed by the initial condition. It also needs a domain. The rule is that the domain is the largest open interval that contains the initial $x$-value and on which the solution and the differential equation are both defined.
Go back to $y = \frac{2}{4 - x^2}$ with $y(0) = \frac{1}{2}$. The formula breaks at $x = 2$ and $x = -2$. The initial point sits at $x = 0$, so the domain is $-2 < x < 2$. The pieces beyond $\pm 2$ do not count, because a solution curve cannot jump across a vertical asymptote.
A second trap comes from the equation itself. For $\frac{dy}{dx} = -\frac{x}{y}$ with $y(0) = 5$, the solution is $y = \sqrt{25 - x^2}$. At $x = \pm 5$, $y = 0$ and the slope $-\frac{x}{y}$ is undefined. So the domain is the open interval $-5 < x < 5$, not the closed one.
5. Exponential growth and decay
Key idea: dy/dt = ky always gives y = y₀e^(kt): starting value in front, rate constant in the exponent, k found from a second data point.
The equation $\frac{dy}{dt} = ky$ says the rate of change is proportional to the current amount. Its solution is $y = y_0e^{kt}$, where $y_0$ is the amount at $t = 0$. If $k > 0$ the quantity grows, and if $k < 0$ it decays. Keep the roles straight. The starting amount is the coefficient in front, and the rate constant sits in the exponent. So $\frac{dy}{dt} = 0.4y$ with $y(0) = 7$ gives $y = 7e^{0.4t}$, never $0.4e^{7t}$.
The relative growth rate $\frac{1}{y}\frac{dy}{dt}$ equals $k$, a constant. That is the signature of exponential change. A quantity growing by a constant amount each year is linear, but one growing by a constant percentage is exponential.
Half-life and doubling time let you find $k$. Suppose 80 grams of an isotope has a half-life of 5 years. Set $40 = 80e^{5k}$, so $e^{5k} = \frac{1}{2}$ and $k = -\frac{\ln 2}{5}$. The model is $y = 80e^{-(\ln 2/5)t}$, which equals $80\left(\frac{1}{2}\right)^{t/5}$. After 15 years, three half-lives have passed, so 10 grams remain.
In context, a question might give $y(0)$ and a second data point, like $y(3) = 50$. Use both: the first fixes $y_0$, and the second gives $k$ through a natural log. Units matter here. If $t$ is in years, $k$ is per year.
6. Newton's Law of Cooling and other models
Key idea: Translate the words into the equation first; for cooling, the rate depends on the gap T − A, so T approaches A over time.
Newton's Law of Cooling says an object's temperature changes at a rate proportional to the difference between its temperature and the room's. Written out, $\frac{dT}{dt} = -k(T - A)$, where $A$ is the ambient temperature and $k > 0$. When the object is hotter than the room, $T - A$ is positive, so $\frac{dT}{dt}$ is negative and it cools. Separating variables gives $T = A + (T_0 - A)e^{-kt}$.
Try it with coffee at 90°C in a 20°C room. The model is $T = 20 + 70e^{-kt}$. If the coffee is 60°C after 5 minutes, then $40 = 70e^{-5k}$, so $k = \frac{1}{5}\ln\frac{7}{4}$. As $t$ grows, $T$ approaches 20 but never reaches it. The coffee cools fastest at the start, because that is when the temperature gap is largest.
Modeling questions test the translation step. "The rate of change of $V$ is proportional to the square root of $V$" becomes $\frac{dV}{dt} = k\sqrt{V}$. "A population grows at a rate jointly proportional to its size and to $1000 - P$" becomes $\frac{dP}{dt} = kP(1000 - P)$.
That last form is logistic growth. It is a BC-only topic, and the AB exam does not test logistic growth as such. An AB item could still give you an equation of this form and ask something routine, like the sign of $\frac{dP}{dt}$ or where $\frac{dP}{dt}$ equals zero. As optional enrichment, the population levels off at the carrying capacity, 1000 here. Its growth rate is greatest at half the carrying capacity, 500, because that is where the product $P(1000 - P)$ is largest.
7. Second-derivative analysis from the differential equation
Key idea: Differentiate dy/dx implicitly, substitute dy/dx back in, and use the sign of d²y/dx² to judge concavity and over- or underestimates.
You can find concavity without solving the equation. Differentiate $\frac{dy}{dx}$ with respect to $x$, remembering that $y$ is a function of $x$. Every time you differentiate a $y$, a factor of $\frac{dy}{dx}$ appears, so you then substitute the original equation for it.
Take $\frac{dy}{dx} = x + y$. Then $\frac{d^2y}{dx^2} = 1 + \frac{dy}{dx} = 1 + x + y$. At the point $(0, 1)$, the slope is 1 and the second derivative is 2, so the solution is increasing and concave up there.
For a product, use the product rule. If $\frac{dy}{dx} = xy$, then $\frac{d^2y}{dx^2} = y + x\frac{dy}{dx} = y + x^2y$. The most common error is treating $y$ as a constant and writing $\frac{d^2y}{dx^2} = y$ from the $x$ term alone.
This analysis pairs with tangent line approximations on free-response questions. Using $\frac{dy}{dx} = x + y$ through $(0, 1)$, the tangent line is $y = 1 + x$, so $y(0.1) \approx 1.1$. The solution is concave up near that point. The curve therefore bends above its tangent line, which makes 1.1 an underestimate. AP exam readers (the people who score free responses) want that reason stated with the sign of the second derivative.
Second derivatives also classify critical points. If $\frac{dy}{dx} = 0$ at a point and $\frac{d^2y}{dx^2} > 0$ there, the solution has a relative minimum.