The new power system needs the collaborative optimization dispatching of “source, network, load and storage integration”. At present, the scheduling automation system adopts relational database to query and store data based on multiple correlation tables, which is difficult to meet the demand of rapid calculation. In this paper, a graph-based calculation method for collaborative intra-day scheduling is proposed. Firstly, the graph database is used to realize the integration of spatial and temporal data of source network load storage; secondly, comprehensively considering the thermal power unit, adjustable load and energy storage; then, the power flow calculation method based on graph calculation is proposed to quickly check the system safety; finally, the system operation status is corrected based on the safety check results until all operation constraints are met. By analyzing the improved examples of IEEE118 and IEEE1354 nodes, the collaborative optimization method of source network, load and storage is proved to improve the computing efficiency.