A new double current closed loop control strategy based on neural network self-adaption is proposed by ana- lyzing the shortages of double current closed loop control strategy. Neural network technology is used in this strategy to realize the self-adaption of controllerg parameters in real-time, and improve operational efficiency of controller, and opti- mize the whole double current closed-loop control structure. The asymmetric fault simulation results show that the control strategy is more effective to DC side voltage fluctuation elimination and voltage ripple component decrease, and it has ...