In order to improve the cruising range of electric bus, this paper studies the deadbeat current prediction vector control system of axial flux permanent magnet synchronous motor (AFPMSM) for electric bus based on the optimal torque distribution method. Firstly, the mathematical model of the three stators-double rotors AFPMSM is established. Secondly, in order to improve the high efficiency range, the efficiency optimal torque distribution method is proposed based on the average torque distribution method and the back propagation (BP) neural network is used to find the optimal torque distribution method. Then a current control strategy based on deadbeat current prediction control is proposed to improve the torque tracking characteristics. Finally, a drive control system is developed for the proposed control strategy, and experimental research and vehicle testing are carried out. The experimental results show that the BP neural network-based torque distribution method designed in this paper increases the high efficiency range of the drive system and improves the cruising range of the electric bus. The drive system using a current controller based on deadbeat current prediction control exhibits good dynamic and steady state performance.