A novel Fault Tolerant Control (FTC) scheme is developed based on Adaptive Dynamic Programming (ADP) for a class of affine nonlinear systems with actuator failures. The estimated actuator failure from a fault observer is utilized to construct an improved performance index function, so the FTC problem can be transformed into an optimal control problem. By using Policy Iteration(PI)algorithm, the Hamilton-Jacobi-Bellman equation can be solved by constructing a critic neural network. Then, the approximate optimal controller can be derived directly. The closed-loop system is guaranteed to be asymptotically stable via the Lyapunov stability theorem. Finally, an example is presented to demonstrate the effectiveness of the proposed method.