Abstract: In order to promote multi energy complementary and low-carbon of data centers, a “source-load” low-carbon optimal scheduling model for data center microgrids is proposesd, which considers power to gas and carbon capture based on stochastic programming. The model considers the constraints of power generation, workloads, electricity to gas system and carbon capture. Carbon capture power plant, electric gas to gas, and combined heat power unit(CCPP-P2G-CHP) are introduced into the power generation side, the captured CO₂ is used as the fuel for the synthesis of methane in the electric gas to electric system. Considering the interactive and batch workloads on the load side, a data center power consumption and workload allocation model are constructed. The simulation results show that the proposed model can minimize the operation cost of the data center microgrid through the optimal scheduling of power generation and workload based on the consideration of carbon emission.