Aiming at the difficulty in selecting the weight matrix of model predictive control (MPC) algorithm in the path tracking control of self-driving vehicles, which leads to low control accuracy and low operating efficiency of the controller, a genetic particle swarm optimization model prediction control (GA-PSO-MPC) algorithm was proposed. Firstly, a vehicle dynamics model was established, the objective function was determined according to the dynamics model and constraints were added to design the MPC controller; Secondly, the genetic particle swarm optimization algorithm (GA-PSO) was used to optimize the weight matrix of the model predictive controller; Finally, a Carsim/Simulink simulation platform was built to compare the tracking performance of GA-PSO-MPC controller with traditional MPC controller, and the simulation of path tracking control under different working conditions with different speeds was completed. The results show that the convergence speed of the controller proposed in this paper after the optimization of the weight matrix by GA-PSO algorithm is improved by 68.85%, and the maximum lateral error is reduced by 63.9%. The operation efficiency and tracking accuracy of the GA-PSO-MPC controller are better than that of the traditional MPC controller at various vehicle speeds, which can effectively solve the problems of low operation efficiency and insufficient tracking accuracy of the traditional model predictive controller.