Abstract: Responding to the Carbon Peaking and Carbon Neutrality Goals and the actual demand for local consumption of distributed renewable energy sources, this paper proposes a method to optimize the capacity of an electricity-hydrogen integrated energy system to meet the demand for multiple loads of electricity, heat, and hydrogen to deal with multi-timescale energy mismatches. Firstly, a hybrid energy storage system model containing batteries and cascade hydrogen storage and a multi-timescale operation strategy are proposed. The model fully exploits the short-term and long-term operational characteristics of hydrogen storage to realize its multi-scale utilization. Secondly, to address the challenges in describing the operation characteristics of long-term energy storage devices based on the traditional typical scenario method, a novel scenario reduction method based on the time-series aggregation method is proposed to provide consecutive scenarios comprising both short-time fluctuations and seasonal characteristics. The technique can effectively speed up the solution and ensure its accuracy. After that, a system capacity optimization model is developed based on the typical scenarios to minimize the life cycle cost and consider factors including energy storage's degradation and the consumption rate of renewable energy. Finally, the effectiveness of the proposed method is verified by cases in Ningbo, Zhejiang, China. The results indicate that the life cycle cost can be reduced by 21.3% with nearly 100% renewable energy consumption compared to the traditional energy management strategy.