An optimized wavelet neural network active disturbance rejection controller (WNN-ADRC) is proposed to solve the nonlinear problems of multiple rocket launcher AC servo system, such as variable load, strong coupling and uncertain disturbance. The mathematical model of the controlled system is obtained by simplifying the current link, and the wavelet neural network (WNN) is embedded into the ADRC for parameter tuning, and the learning algorithm of the WNN is optimized by using the method of hierarchically adjusting the learning rate to obtain the WNN-ADRC. The WNN-ADRC is used to control the servo system of the rocket launcher to realize the accurate estimation and compensation of the nonlinear characteristics. The numerical simulation results show that the WNN-ADRC can improve the static response and dynamic performance of the servo system, and has the advantages of fast response and high control precision compared with the traditional ADRC.