To solve the problems of external disturbance and non-linear characteristics in the positioning control of the servo system of antiaircraft gun, an improved active disturbance rejection controller (WNN-ADRC) based on wavelet neural network is proposed. The Levenberg-marquardt algorithm is used to optimize the wavelet neural network. Using wavelet neural network adjust the error correction gain coefficient in the expanded state observer on-line, design an active disturbance rejection controller based on wavelet neural network to achieve accurate estimation and compensation of nonlinear characteristics, and verified by simulation experiments. Simulation results show that this control strategy makes the system has better steady-state performance and strong anti-interference ability.