Abstract: In response to the severe decarbonization challenges faced by high-energy-consuming data centers and the technical bottlenecks of green energy supply, this paper proposes a collaborative optimization control strategy based on a shared hydrogen energy storage power station. By constructing a multi-energy complementary coupling system of "electricity-electricity", "electricity-hydrogen", "hydrogen-hydrogen" and "hydrogen-electricity", a multi-objective dynamic balance mechanism characterized by "low hydrogen production cost, high peak-shaving revenue, and strong backup guarantee" is established, which is combined with the Alternating Direction Method of Multipliers （ADMM） to achieve the global electricity-hydrogen interaction optimization under privacy protection. Case studies show that the proposed multi-energy supply scheme can stably maintain the green electricity consumption rate of the data center above 98.1%. Based on the cross-seasonal energy transfer characteristics of hydrogen energy storage, hydrogen power generation also provides a stable peak regulation capacity in extreme scenarios, and the peak regulation revenue reaches 1.8104 million yuan per year. Compared with the backup scheme of diesel generators, the total cost savings of the hydrogen fuel cell coupled with electric energy storage over a ten-year period reach 14.749 million yuan, and the carbon emission intensity in the whole life cycle is reduced by 96.8%. This study provides a quantitative method and engineering reference for solving the collaborative problem of "high reliability-low carbonization-low cost" in data centers.