Abstract: This paper proposes a fault detection and fault tolerant control method for a class of nonlinear systems based on neural networks. In the approach, a fault-estimator is designed using neural networks for on-line estimation of faults, then the fault detection is achieved. Once a fault is detected, a compensating controller is introduced into the post-fault systems to accommodate the faults. The stability analysis is also given based on Lyapunov method. Finally, the proposed method is used in fault detection and fault tolerant control of nonlinear DC-motor, simulation results demonstrate the effectiveness of the proposed method.