The dynamic response of the pitch angle will be slowed down when the pitch actuator of wind turbines fails, and the fault will lead to the output power fluctuations. In the light of the identification theory and the analytical theory, a fault tolerant control method combined with the fault estimation is proposed to resolve the problem. The second order transfer function is transformed into the identification equation by using the Euler transformation. Next, the time varying natural frequency and damping ratio of pitch actuators are estimated using the sliding data window least squares based iteration (SDW-LSI) identification algorithm. According to the theory of analytics, a compensation equation is derived, and the values estimated are fed back to the compensation module to adjust the relations before and after faults occur in the pitch actuator to solve the problems of slow pitch angle and the output power fluctuation. Finally, the feasibility of the algorithm is validated by choosing the fault of high air content in hydraulic oil. This work is supported by National Natural Science Foundation of China (No. 61572237 and No. 61573167).