Conventional model predictive control method of permanent magnet synchronous motor (PMSM) suffers from high torque ripple and speed fluctuation, which affects the ride comfort of vehicle. A model predictive control method based on state transition constraint of a PMSM is proposed. First, the state transition probability matrix is calculated based on the historical data of the switching state of the PMSM. Secondly, the constraint error is obtained according to the current switch state and state transition probability matrix. This constraint error can limit the switching state of the switch at the next moment. Then, the cost function including the state transition constraint error is formulated in the model predictive control algorithm. The optimal switching state at the next moment is searched according to the cost function. The results of simulation study and experiment show that the improved model predictive control strategy proposed in this paper has better torque and speed response characteristics in terms of speed ripple and torque fluctuation. The results show that the proposed model predictive control method based on state transition can be used in the control of vehicle PMSM. It is essential to improve the ride comfort of the vehicle.