The purpose of this paper is to design an intelligent wheelchair that can follow the patient's walk independently. Besides, it can help rehabilitation patients recover better through the intelligent automatic control system. On the one hand, the cost compensation function is introduced to improve the traditional Astar algorithm. It can make the intelligent wheelchair more adaptable to the environment. The underground environment is simulated to verify it. Simulation shows that the surface intelligent wheelchair can identify the path more accurately. On the other hand, the wheels of the intelligent wheelchair and the driving motor of the mechanical arm both use IPMSM. To control the motor more efficiently and accurately, the MTPA control method is adopted in this paper. The process of an intelligent wheelchair starting, moving at a constant speed, and slowing down and stopping on flat ground is simulated and verified. The simulation shows that the intelligent wheelchair can realize the start and stop acceleration which makes the patient feel comfortable. The results show that the intelligent wheelchair has a good effect on the recovery of patients.