In the dynamic wireless power supply system for electric vehicles, the output current is influenced by factors, including primary rail current, relative position between pick-up unit and track, capacity of battery packs. Negative effect of those factors may lead to unsteady output. Aimed at the question of accurately modeling and output control caused by higher-order switch nonlinear behavior and multi-disturbance factor. An output current-stabilizing control strategy based on back-propagation (BP) neural network is proposed, making heavy use of nonlinear function approximation character and powerful generalization capability for neural network. With this strategy, the system was robust enough to unknown disturbance and parameter variation and output current can keep constant unswervingly. Finally, simulation results show that this control strategy has obvious advantages both in overshoot and setting time over conventional PID control method.