Abstract: In view of the volatility of the wind farm output and the inaccuracy of the wind prediction, a digital twin hybrid energy storage (DTHES) based control strategy is proposed to optimize the operating efficiency of the energy storage equipment. The strategy considers the fluctuation of the wind power grid connection, the error of the prediction results and the efficiency of the energy storage equipment. Firstly, the Attention-GRU prediction model is established by introducing the Attention mechanism into the gated recurrent unit. Then, combined with the grid-connected power fluctuation standard and the response characteristics of different energy storage devices, the virtual variational mode decomposition (VVMD) is performed on the digital twin prediction data to obtain the primary distributed power. Finally, considering the limitations of the operating efficiency and service life of the energy storage equipment, the hybrid energy storage service system is used to modify the primary distributed power of the virtual hybrid energy storage to obtain the final distributed power. Experiments based on the wind power plant data in a certain area of Northeast China have verified that the strategy can improve the efficiency and accuracy of the wind power prediction, suppress the fluctuation of the wind power grid connection effectively, and enhance the utilization efficiency of the energy storage equipment.