Aiming at the problem of model predictive controller(MPC) oscillations caused by excessive position deviation and constraints in the early stage of tracking tasks for unmanned surface vehicle(USV), the MPC control strategy guided by virtual USV was proposed, which used virtual transition trajectories to replace the target trajectories of actual USV, and conversion conditions were designed to determine when the USV  exits the virtual guidance strategy during execution. To solve the stability problem of MPC, based on the quasi infinite model predictive control theory, a terminal penalty function was added in the MPC design, and by constructing a Lyapunov function, the stability of the proposed method in the finite time domain was demonstrated. A nonlinear disturbance observer was introduced to observe the disturbances in complex marine environments, and was compensated by using controller. The simulation experiments of circular, sinusoidal, and straight working conditions have verified the effectiveness and accuracy of the proposed method, which can achieve trajectory tracking control of USV in complex marine environments.