Abstract: The existing transmission channels in high load density areas are limited and the proportion of flexible power sources in the system is low, so the adjustment ability to deal with the random fluctuation of renewable energy at the source end and load demand at the load end is insufficient. As an efficient flexible resource, energy storage system (ESS) has the advantages of fast response and two-way power conversion, which can provide flexible support for power system. Based on the bi-level programming model, this paper establishes an energy storage optimization configuration method to improve the overall and local flexibility of high-density load areas. The upper model makes strategic decisions on the configuration scheme of energy storage. The lower layer model simulates the operation of typical daily scenarios based on scenario analysis and uses CVaR to quantify the risk cost of insufficient flexibility. The model fully considers the balance of power supply and demand, new energy consumption and the supporting role of grid structure. The double-layer model is transformed into a single-layer model by adding association constraints. Finally, the commercial software package GUROBI is called through the MATLAB platform to complete the mixed integer programming optimization, and the optimal configuration scheme is obtained. The effectiveness of the configuration method is verified in a high load density area of an actual system.