The trajectory tracking of autonomous vehicles is closely related to vehicle motion control, which determines whether the vehicle can effectively follow the planning instructions of the decision-making layer and correctly output the control instructions of the drive and steering actuators according to the dynamic characteristics of the vehicle. A two-degree-of-freedom vehicle dynamics model was established, and a model predictive control architecture based on vertical and horizontal coupling dynamics was proposed by using the adaptability of Model Predictive Control (MPC) in a complex nonlinear and multi-constraint environment, in which the upper-level controller accurately adjusted the longitudinal speed and acceleration to ensure that the vehicle ran along the expected speed trajectory. The lower controller tracks the lateral trajectory and realizes the efficient tracking and control of the planned path of the vehicle by accurately controlling the front wheel rotation angle. The common obstacle bypass scenarios of automobiles are set up for simulation tests, and the proposed strategy is compared with the classical Proportional-Integral-Derivative (PID) and Linear Quadratic Regulator (LQR) controls, and the real vehicle experiments are carried out. The results show that the proposed strategy shows more stable and accurate control performance in different scenarios, which can improve the trajectory tracking accuracy of autonomous vehicles, optimize the control smoothness and improve the ability of vehicles to adapt to complex traffic scenarios.