Aiming at the problem that the firing accuracy and stability of shipborne rocket launcher will be reduced due to the influence of wind and waves and its own disturbance in the launching process, a RBF neural network sliding mode-fuzzy PID control method is proposed. The RBF neural network is used to weaken the chattering of sliding mode and improve the response speed and robustness when the error is small. The fuzzy rules are used to adjust the PID parameters to improve the control accuracy when the error is large. The simulation results show that the hybrid control strategy can make the AC servo system of shipborne rocket launcher have higher firing accuracy and reaction speed, and improve the system performance.