Conventional PID and fuzzy algorithms have the problems of strong coupling and nonlinearity in the brushless DC motor (BLDCM) control system, and signal distortion is prone to appear under the action of interference. In order to solve these problems, the taxi light servo steering system is taken as the application background, the BLDCM three-closed-loop control system is taken as the research object, BP neural network, fuzzy control and PID algorithm are combined, and a kind of PID control based on fuzzy parameter modified BP neural network is proposed. Through Simulink modeling and simulation, the dynamic response characteristics of this strategy and conventional control algorithms under torque disturbance and magnetic flux disturbance are compared. The simulation results show that this improved control algorithm has excellent performance in the BLDCM position control system.