Abstract: The widespread construction of data centers is a developing trend in the intelligent era. The rational planning of data centers, considering their load transfer characteristics while ensuring service quality, holds great significance in reducing the overall cost of data center planning and enhancing the flexibility of new power systems. Therefore, a data center computing power and electric power resource optimization planning method considering the guidance of multiple electricity markets is proposed. Firstly, a virtual generator and a virtual energy storage model are established using the polyhedral bounding contraction method to describe the aggregated characteristics of the data center park. Then, based on the settlement rules of different electricity markets, the methods and economic benefits of data center parks participating in the electricity frequency regulation-capacity market are investigated. On this basis, with the goal of minimizing comprehensive costs, the construction capacity of generators, energy storage, and the number of servers in the data center park are optimized. Finally, taking Alibaba’s publicly available operational data as an example, it is demonstrated that considering diversified power markets can greatly reduce the planning costs of data centers, incentivize investment and construction of more flexible resources in data center parks, and provide effective frequency regulation services and capacity support for the power grid.