Given the vision of large-scale deployment of a grid-connected distributed Energy Storage System (ESS) in the distribution network, it is necessary to study the capacity optimization of ESS for striking a balance among investment, Renewable Energy (RE) consumption, power supply security and price arbitrage. In this paper, an ESS partition model based on an improved flame propagation model is proposed. The results of ESS partition are obtained by constructing indices such as the flammability of nodes, the wind direction of flame propagation, and the speed of flame propagation. Then, an optimization method for ESS capacity in partition is established. This is to maximize the annual net profit and an integer programming method is used to solve the model. Finally, the effectiveness and feasibility of the method are verified by conducting case studies on the IEEE 33 bus distribution network system. The case study results show that the rational configuration of ESS capacity can improve the economy of planning, and the proposed method can improve Photovoltaic (PV) local consumption ability and power supply reliability of the distribution network. This work is supported by National Natural Science Foundation of China (No. 51707099), Science and Technology Project of State Grid Jiangsu Electric Power Co., Ltd. (No. J2019055), and Undergraduate Scientific and Technological Innovation Project of Jiangsu Province (No. KYCX19_0801).