In order to study the possibility of network reconfiguration and cluster division to improve distributed generation configuration in the distribution network, a two-layer model of dynamic reconfiguration and cluster division is proposed, which can obtain the best reconfiguration strategy and cluster division at the same time. The distributed photovoltaic (DPV) and energy storage system (ESS) capacity and location optimization experiments are carried out through 4 schemes, the optimization indicators such as DPV capacity and annual overall cost and operation indicators such as network loss and voltage level are compared. The results show that under the DPV and ESS capacity and location optimization scheme based on the two-layer division result, the DPV capacity is the largest and the annual overall cost is the lowest. The examples show that the bidirectional dynamic reconfiguration strategy can greatly expand the room for improvement of DPV consumption while reducing planning costs. It has high reference value when solving high-proportion distributed power optimization problems. This work is supported by the National Natural Science Foundation of China (No. 61773051 and No. 61963026).