Because of the many advantages of PMSM, a model predictive control method of PMSM based on maximum torque current ratio (MTPA) control to maximize the utilization of reluctance torque was presented. The typical problems of several optimization targets were studied in detail, including switch state constraints, MTPA optimization, maximum current limit and delay compensation strategy, using discrete digital processing techniques. The simulation results of MATLAB/Simulink showed that the control system achieved the maximum torquecurrent ratio control and the dynamic response process was very fast, and the working efficiency of the PMSM model predictive control system was improved remarkably, and the fluctuation of the motor parameter showed strong robustness.