Abstract: A novel active fault-tolerant control strategy for wind energy conversion system is proposed to deal with part failure of actuator. According to the state observer value, a self-adaptive radial basis function(RBF) neural network is introduced to reconstruct the actuator fault on-line. The switch gain of the sliding fault-tolerant controller is designed based on the reconstructed fault. Then a combination of both on-line fault diagnosis and tolerant control for the system is realized. Therefore, the system stability can be proven. Finally, the simulation results show that the system power coefficient and tip speed ratio can be maintained at the optimal value under the rated wind speed even if the actuator has failed, and the maximum wind energy capture can be realized.