Extreme heat is increasingly threatening urban sustainability under climate change and rapid urbanization. Here, we develop a heat resilience index integrating heat intensity and recovery duration and evaluate how urban scale and land-cover changes have shaped heat resilience across 70 major Chinese cities between 1990 and 2024. Using 1 km ERA5-derived temperature data and urban indicators, this study combines machine learning, geographically weighted regression, and Shapley additive explanations (SHAP) analyses to reveal spatially heterogeneous urbanization-heat resilience relationships. Results show that heat resilience deteriorated by nearly 120% during early urbanization stages before stabilizing. Increasing built-up area (+3 × 108 m2) or total building height (+7 × 104 m) corresponds to a 3.6 × 105 K·s increase in heat resilience index, indicating deteriorated resilience, whereas green area expansion (+1.6 × 109 m2) is associated with a 3.6 × 105 K·s decrease, reflecting improved resilience. Built-up density and green coverage emerged as the dominant drivers, showing positive and negative associations with the heat resilience index, respectively. These findings provide actionable insights for climate-resilient urban planning under escalating heat extremes. Broader context: Rapid urbanization and intensifying heatwaves are making cities more vulnerable to extreme heat, with growing consequences for public health, infrastructure, energy demand, and social equity. By tracing heat resilience across 70 Chinese cities over more than three decades, this study shows how urban growth and land-cover change can either weaken or improve a city’s ability to cope with heat extremes. The work bridges climate science, urban planning, remote sensing, and artificial intelligence to provide a clearer basis for action.In the longer term, these findings can support heat-adaptive design, greener development pathways, and more effective planning and policy decisions for resilient and livable cities. Although developed using Chinese cities as case studies, the framework can be adapted to cities with different climatic and development conditions, providing opportunities to support climate adaptation and sustainable urban development in diverse urban contexts worldwide.