Abstract: Applying the new generation of artificial intelligence in smart grid and Energy Internet, to achieve high proportion renewable energy access to the power grid in a timely and effective manner, the deep deterministic policy gradient（DDPG）algorithm based on deep learning is applied in the optimized operation of active distribution network（ADN）. Firstly,DDPG return function of optimization model for ADN with multiple microgrids is constructed, which can minimize the total node voltage deviation and line loss of ADN. The proposed function can also minimize the variation of the power regulation of microgrid to reduce the impact on operation of the microgrid, and maintain the balance of tie-line power blance to reduce the impact on the distribution network. Secondly, DDPG sample data processing, design of return function, model training and learning process of optimization control for ADN are analyzed. Finally, the effectiveness of the algorithm is verified by the improved IEEE 14-bus example simulation.