Configuring energy storage power stations is an effective measure to alleviate the randomness and volatility of renewable energy generation. Considering the randomness of renewable energy and the optimization goals of grid diversification, energy storage planning techniques become a crucial issue in grid optimization and layout. In this paper, a multi-objective optimization strategy for energy storage configuration in a grid considering the randomness of renewable energy is proposed. Firstly, the k-means algorithm is used to extract monthly characteristic information of renewable energy generation, forming a typical daily dataset. Then, an optimal power flow model is established, and the optimal energy storage configuration is solved based on second-order cone optimization. This method reduces the impact of randomness in renewable energy generation on energy storage planning and configuration. Meanwhile, considering variable weights, it achieves comprehensive energy storage planning considering the optimization goals of power generation, load, and energy storage. This research can provide references for subsequent studies on multi-objective planning and configuration techniques for energy storage power stations.