Abstract: S surface controller is proved to be effective in application to the motion control of autonomous underwater vehicle (AUV). However, it is difficult to adjust the control parameters manually. Choosing the optimum parameters for the controller of a certain AUV is an important problem. Thus, a modified particle swarm optimization (MPSO) algorithm based on immune theory and nonlinear digression strategy of inertia weight was proposed to optimize the control parameters of a certain AUV. A semiphysical simulation system for AUV was presented as a platform to verify the proposed control method. And its structure was described. The simulation results indicate that the semiphysical simulation platform is helpful, the optimization algorithm has good local and global searching abilities, and the control method is feasible in application to AUV.