Abstract: To cope with the stochasticity and uncertainty of large-scale offshore wind power, energy storage as a flexible resource receives increasingly widespread attention. However, the existing research on energy storage configuration for grid-connected new energy sources focuses on the power side, and lacks research on the optimal allocation of grid-side energy storage. To address this issue, a two-stage grid-side energy storage configuration method is proposed based on the consideration of the prediction error of offshore wind power in different output intervals. In the first phase, based on offshore wind power output and load data, a multi-scenario stochastic planning method is used to optimally configure the energy storage capacity in an economical way, to maximally compensate for the wind power prediction error and to reduce the amount of wind abandonment and load shedding. In the second stage, a multi-temporal optimal tidal current model based on preferred siting is established, taking system security and network architecture into account, and the minimum unit cost is taken as the objective for energy storage siting and capacity setting. Finally, the modified IEEE39 node system is used for simulation, which is combined with the comprehensive evaluation model of energy storage to verify the effectiveness of the method in this paper.