With the introduction of the “dual carbon” targets, building a clean and efficient power and energy system has become necessary. However, while the large-scale grid-connected renewable energy such as wind and solar power alleviates the energy and environmental issues, the strong uncertainty of their output also leads to significant wind and solar power curtailment. Reasonable sizing of energy storage is crucial to improve the utilization of renewable energy and whole system economics. This paper proposes an optimal configuration method for hybrid energy storage considering multi-energy coordination and source-load uncertainty. First, a probability distance-based fast forward selection method reduces scenarios to represent uncertainty characterization. Then,from the perspective of optimal system economics, a bi-level stochastic optimization model integrating energy storage planning and multi-energy operation is established: the upper-level planning model determines optimal capacity configuration for battery and pumped hydroelectric storage, while the lower-level operational model coordinates the multi-timescale dispatch of wind, solar, and storage resources. The model was solved using an iterative strategy to determine optimal capacity configuration of battery and pumped hydroelectric energy storage considering multiple scenarios. Finally, the proposed method was validated through case studies on the Zhejiang power system to demonstrate its effectiveness in improving system economic efficiency and renewable energy utilization.