The traditional model predictive control for dual permanent magnet synchronous motor system driven by fiveleg inverter reduced the control accuracy of the two motors due to the single vector action in one control cycle. Aiming at this problem, an optimization model predictive control strategy was proposed. By distributing the duty cycles of the active voltage vector and the zero voltage vector of the two motors, the torque ripple was reduced to improve the control precision of the two motors, and then the control performance of the motor system was improved while satisfying the control independence of the two threephase motors. The feasibility and effectiveness of the proposed optimal model prediction control were verified by simulation results.