In order to solve the problem that the driving condition of the electric vehicle (EV) cannot be covered by the high efficiency area of traditional single motor drive system, an optimal allocation strategy for the multi-motor drive system of EV was presented based on model predictive control. Firstly, taking the whole multi-motor drive system as research object, the model of motor and longitudinal dynamic model of automobile were established. The method to improve the efficiency of the vehicle with different front and rear axle motors in the high efficiency area was discussed. Secondly, the efficiency of the motor at specific speed and torque operation points was calibrated by the bench test. It was found that the high efficiency ranges of the two motors covered the working operation points of high speed cruising and low speed startup conditions, respectively. By introducing the driving force distribution ratio of front and rear motors, the efficiency diagram of the two motors was transformed into the vehicle driving efficiency diagram. Thirdly, theoretical analysis and simulation verification of permanent magnet synchronous motor (PMSM) system with model predictive torque control were carried out, showing that the method can ensure the rapid response of torque. Finally, the validity of the range improvement and vehicle efficiency were verified by the hardware in the loop (HIL) experiments. The results showed that the driving force distribution can integrate the high efficiency range of two motors and improve the efficiency of vehicle. Moreover, the model predictive torque control (MPTC) based on the motor benchs parameters can suppress the torque ripple in the low speed range and improve the accuracy of the output torque. The results can provide theoretical basis for the design of distributed driving system.