In conventional direct torque control (DTC), the switching instant of the voltage vector is variable and the switching point of voltage vector is limited by the hysteresis band. The model predictive control (MPC) is applied in permanent magnet synchronous motor (PMSM) DTC system, and an optimized voltage vector prediction selection method is analyzed. On one hand, the predicted stator flux linkage and electromagnetic torque are input into the corresponding hysteresis controller, and the appropriate voltage vector is selected from the limited set of voltage vectors to ensure that the predicted flux linkage and torque are within the corresponding respective hysteresis. On the other hand, the cost function and optimization strategy are introduced to select the corresponding optimal voltage vector under different conditions, and the control objectives besides current protection can be considered at the same time. The proposed algorithm is experimentally verified on a salient pole PMSM drive platform. Experimental results show that the proposed control strategy can effectively control the torque within the hysteresis controller, which verifies the effectiveness of the strategy.