<p>This paper proposes a dispatching strategy of hybrid energy storage considering power forecast errors to address renewable energy fluctuations and user demand in power systems without conventional power sources. To improve the system benefits, a two-stage scale dispatching model for wind-solar hybrid systems that incorporates power forecast error compensation has been devised. This comprises a day-ahead dispatching model and a power allocation model. The objective of the day-ahead dispatching model is to maximize benefits by optimizing the output of the system. The benefits, costs and constraints of the energy storage system (HESS) are considered in the model. For the energy storage unit, ICEEMDAN is used to decompose the reference power of the HESS and delineate the high and low frequency cut-off points by the energy entropy difference. This is used to determine the optimum capacity configuration and power output of the battery and flywheel. Simulation results show that the proposed method reduces the rate of power loss from 27.49 % to 6.01 % on a typical day. The optimized system achieves efficient utilization of renewable energy, ensuring the stable operation of the microgrid. The use of the ICEEMDAN and energy entropy to decompose the HESS reference power, avoiding modal aliasing problems and providing theoretical support for power allocation.</p>