Abstract: To improve the accuracy and economic efficiency of reserve capacity allocation in high-penetration renewable energy power systems, enhance renewable energy absorption, and optimize system operational economics, this study first employs the generalized convex hull uncertainty set method to effectively describe the spatiotemporal correlation of source-load output, thereby accurately characterizing the essential reserve capacity demand under source-load uncertainty conditions. Subsequently, considering the flexibility reserve supply potential of the power system's sources, loads, and storage, a system reserve decision-making mechanism is proposed, accounting for the flexible regulation of multiple types of reserve resources, to achieve the refined allocation of reserve resources and ensure the system's safe and stable operation under various scenarios. Finally, simulation analysis is conducted based on a two-area, 28-node interconnected system, and the simulation results validate the effectiveness of the proposed method.