<p>Flooding remains one of the most destructive and recurrent natural disasters, especially in large river basins affected by climate change and rapid urbanization. While prior studies have mapped flood-prone areas, the mechanisms by which flood risks propagate across regions remain poorly understood. This study introduces a novel spatial framework to identify potential flood risk (PFR) diffusion corridors and their interregional transmission mechanisms in the upper Yangtze River Basin, China. We integrated the InVEST-SWY hydrological model with extreme value theory to simulate monthly quickflow (QF) and delineate flood-prone periods and zones. Morphological spatial pattern analysis (MSPA) was used to identify core flood risk sources, while a comprehensive resistance surface—composed of six environmental factors—was constructed using the Analytic Hierarchy Process. Circuit theory and the Minimum Cumulative Resistance model were then applied to simulate PFR connectivity networks and detect critical corridors and hotspots. The results revealed that May to July represents the peak risk period, during which 25, 21, and 17 PFR hotspots were identified, respectively. The number of risk corridors decreased from 104 in May to 53 in July, indicating growing spatial concentration and network compaction. Most PFR clusters were concentrated in the central and southeastern sub-regions, characterized by low resistance and high urbanization. This framework enhances the understanding of flood risk connectivity and offers actionable insights for prioritizing mitigation efforts. By identifying critical transmission nodes and corridors, the study supports sustainable watershed planning and targeted flood resilience strategies.</p>