Heat exposure drives substantial global disease and mortality burdens, making spatially explicit heat health risk assessment essential for targeted adaptation. Existing studies apply diverse modeling approaches yet rarely evaluate them systematically because reliable epidemiological validation data remain scarce; conventional risk indices further depend on subjective indicator aggregation and expert weighting. Here we show an integrative geographically neural network weighted regression (GNNWR) framework validated against epidemiologically derived heat-attributable mortality fraction at the subdistrict level. The framework achieves stronger predictive performance and better generalization than ordinary least squares, multiplicative model, random forest, geographically weighted regression (GWR), multi-scale GWR, and other geographically weighted machine learning models, with consistent alignment between training and test sets and no evidence of overfitting. By learning spatial weights through a neural network, it simultaneously captures spatial non-stationarity and nonlinear relationships among meteorological, air pollution, demographic, socioeconomic, and built environment variables without requiring predefined index construction. It enables prediction in data-sparse areas and generates consistent high-resolution (100 m grid), subdistrict, and district-level risk maps, while identifying fine particulate matter (PM_2.5) as the second strongest risk factor associated with heat-related mortality risk, after temperature. This validated, data-driven approach provides a transferable methodological foundation for identifying high-risk populations and informing precise public health interventions and urban climate adaptation strategies.

Green-blue spaces (GBS) are pivotal in mitigating thermal discomfort. However, their management lacks guidelines rooted in epidemiological evidence for specific planning and design. Here we show how various GBS types modify the link between non-optimal temperatures and cardiovascular mortality across different thermal extremes. We merged fine-scale population density and GBS data to create novel GBS exposure index. A case time series approach was employed to analyse temperature-cardiovascular mortality association and the effect modifications of type-specific GBSs across 1085 subdistricts in south-eastern China. Our findings indicate that both green and blue spaces may significantly reduce high-temperature-related cardiovascular mortality risks (e.g., for low (5%) vs. high (95%) level of overall green spaces at 99th vs. minimum mortality temperature (MMT), Ratio of relative risk (RRR) = 1.14 (95% CI: 1.07, 1.21); for overall blue spaces, RRR = 1.20 (95% CI: 1.12, 1.29)), while specific blue space types offer protection against cold temperatures (e.g., for the rivers at 1st vs MMT, RRR = 1.17 (95% CI: 1.07, 1.28)). Notably, forests, parks, nature reserves, street greenery, and lakes are linked with lower heat-related cardiovascular mortality, whereas rivers and coasts mitigate cold-related cardiovascular mortality. Blue spaces provide greater benefits than green spaces. The severity of temperature extremes further amplifies GBS's protective effects. This study enhances our understanding of how type-specific GBS influences health risks associated with non-optimal temperatures, offering valuable insights for integrating GBS into climate adaptation strategies for maximal health benefits.