Abstract: A control method combining the MPC（Model Predictive Control） and MPSO（Modified Particle Swarm Optimization） is developed for path following of underactuated surface vessels and avoidance of static or dynamic obstacles. The combination of LOS（Line-Of-Sight） guidance and ILOS（Integral Line-Of-Sight） guidance is used to solve the problem of underactuated system and to compensate the steady state cross-track error caused by drift. A nonlinear optimization problem is formulated to minimize the deviation of actual heading from desired heading and the rate of change of rudder angle. NMPC（Nonlinear Mode Predictive Control） method is used to realize heading control, meanwhile, the collision avoidance factors are embedded in the path following control problem as a nonlinear penalty term. The MPSO algorithm is used to solve the nonlinear optimization problem and generate control commands. Simulation experiments under the condition of static and dynamic obstacles are carried out with the ONRT（Office of Naval Research Tumblehome） model, and the results are compared to those from LPSO（Linear Particle Swarm Optimization） algorithm to verify the effectiveness of the proposed method.