Global climate change and rapid urbanization have intensified flood risks worldwide, especially in cross-regional watersheds where jurisdictions often implement mitigation strategies independently. Although grey infrastructure is widely used to address these heightened risks, its fragmented application frequently shifts hazards to adjacent regions and causes adverse ecological impacts. In contrast, green infrastructure (GI), an interconnected network of natural and semi-natural areas, offers a promising nature-based solution, yet variability in terrain, soils, land use, and hydrological connectivity complicates the development of universal GI planning guidelines. Thus, this study addresses two critical questions: (1) How do changes in flood risk management performance (FRMP) in one region affect neighboring regions? (2) How can GI planning be tailored to watershed heterogeneity? Focusing on three contiguous regions in China's Yangtze River Delta Ecological Green Integration Demonstration Zone, we simulated flood processes using the SCS-MIKE11 hydrological-hydrodynamic model, optimized GI spatial configurations via Simulated Annealing, and applied the TOPSIS to select configurations that balance FRMP across all regions. Results show: (1) significant interregional FRMP correlations, with midstream negatively correlated with upstream (p <0.001) and downstream (p <0.001); (2) dispersed GI spatial configurations better accommodate watershed heterogeneity; (3) prioritizing FRMP at regional boundaries when configuring GI effectively mitigates watershed-wide flood risks; (4) distributive justice, integrated land and water management are essential for cross‐regional flood challenges. This study reveals interregional FRMP coupling and pioneers a heterogeneity-responsive GI optimization, offering planners a novel decision-support tool for coordinated GI planning for cross-regional flood risk management.