Considering the coordinated optimization of generation, network, load and storage, this paper builds a three-layer planning model of the active distribution network frame by taking the minimum annual comprehensive cost as the goal. Through the mutual transfer and coordination of the three-layer model, the distributed power supply, the interruptible load, the distribution network frame and the battery energy storage are optimized coordinately in the distribution network. The hybrid algorithm combining the modified genetic algorithm with the tree structure coding and the original dual interior point method is used to solve the model. The calculation results of the system with 29 nodes verifies the correctness and effectiveness of the proposed method which also indicate it can effectively reduce the cost of distribution network planning and improve the load characteristics of distribution network by considering demand side management measures and the optimization of the output of battery energy storage.