PV integrated 5G base stations can effectively reduce the energy cost of communication operators, but the energy consumption mode of 5G base station with distributed PV can affect the stable operation of the active distribution network, resulting the demand of energy storage configuration for both communication operators and the active distribution network. For security and economic considerations, this paper introduces the energy storage operator as a third party to build a shared energy storage system. A multi-entity shared energy storage optimization configuration method considering the energy consumption mode of PV integrated 5G base stations is proposed, and the optimization configuration problem is constructed as a bi-level mixed integer planning model. The upper model aims to optimize the cost of shared energy storage system configuration to ensure economy; the lower model aims to optimize the net load shaving and filling effect of the active distribution network and the total operating cost of large-scale 5G base stations, aiming to reduce the load peak-to-valley difference while promoting the consumption of renewable energy. Finally, the reformulation and decomposition algorithm is adopted to solve it. The simulation results show that the proposed shared energy storage optimal configuration method can achieve peak shaving and valley filling in the active distribution network and win-win effect among multi-entities.