An optimal tracking controller based on the backstepping method is designed for a complex system with input constraints. The control is divided into two parts， steady-state tracking control and transient optimal control. First， the finite-time-convergent differentiator is introduced to obtain the time derivatives of virtual control laws to avoid the problem of “explosion of terms” within the backstepping structure. Considering the uncertainty in the system， a nonlinear disturbance observer is designed to ensure the robustness of the controller. Second， an auxiliary error compensation strategy correct the control input. To fulfil the optimal control， the adaptive dynamic programming is used to design the transient optimal controller. Finally， the effectiveness of the control method are proved by simulation results.