<p>The Urban Heat Island (UHI) effect, characterized by higher temperatures in cities than rural areas, impacts the environment and the quality of life of city inhabitants. This study examines the spatiotemporal dynamics of the UHI effects in the megacity of Dhaka using Google Earth Engine remote sensing data and climate records. High-resolution Land Surface Temperature (LST) derived from Landsat-4/5 (TM), 7 (ETM+), and 8 (OLI_TIRS) images spanning from 1990 to 2023 were analyzed to determine thermal trends across the Dhaka Metropolitan area. To evaluate the effects and effectiveness of UHI, various mitigation strategies, such as green roofs, urban forests, water conservation, high-albedo building materials, proper land-use planning, and population management, were considered. Normalized Difference Vegetation Index (NDVI), Normalized Difference Built-up Index (NDBI), Urban Index (UI), Normalized Difference Water Index (NDWI), and Albedo were estimated for different years, which were used for developing machine learning (ML) models to estimate LST. The relative performance of different ML models was examined to select the effective one for quantifying the impacts of mitigation strategies on reducing LST. The results showed a significant inverse relationship between LST and albedo, NDVI, and NDWI, and a direct relationship with UI and NDBI. The Support Vector Machine (SVM) model was the most effective in simulating LST compared to other ML models. Application of the SVM model revealed that a 10% reduction in NDBI could decrease LST by 15%. Increasing NDWI and NDVI by 20% can lower LST by 7% and 12%, respectively. While challenging, a 25% increase in NDVI has the potential to reduce LST by 15%. Although data and model uncertainties exist, this study offers pragmatic interventions to the planners and policymakers to significantly reduce UHI effects, improve living conditions, and build resilience against potential climate impacts in the rapidly growing Dhaka.</p> Graphical Abstract <p>This graphical abstract visually summarizes the methodology, major findings, and proposed mitigation strategies for controlling the Urban Heat Island (UHI) effect in Greater Dhaka City. The upper section highlights the study’s use of multi-temporal remote sensing data from Landsat satellites (TM, ETM+, and OLI/TIRS) spanning the period from 1990 to 2023. Key environmental indices—NDVI, NDBI, NDWI, Urban Index, and Albedo—were extracted using Google Earth Engine to understand urbanization-driven changes in land surface temperature (LST). The middle section presents the machine learning modeling process, where the Support Vector Machine (SVM) was identified as the best-performing model to predict LST patterns. Sensitivity analyses showed that built-up areas (high NDBI) are major contributors to elevating LST, while green spaces (high NDVI) and water bodies (high NDWI) are critical for cooling. Quantitative findings demonstrated that a 10% decrease in NDBI could lower LST by 15%, while a 20% increase in NDWI and NDVI could reduce LST by 7% and 12%, respectively. The lower section illustrates practical mitigation strategies recommended based on model results. These include enhancing green infrastructure (urban forestry, green roofs, vertical gardens), promoting water-sensitive urban design (conserving rivers, lakes, and wetlands), applying cool materials (high-albedo pavements and reflective roofs), and improving urban planning (integrated open spaces and climate-responsive land use management). Community engagement and sustainable transport initiatives are also emphasized as essential behavioral adaptations. Overall, the graphical abstract captures the comprehensive approach of the study, from historical LST trend analysis and machine learning-driven sensitivity evaluation to proposing localized, practical, and scalable solutions. It highlights how targeted environmental management can mitigate UHI intensity, enhance urban livability, and contribute to the climate resilience of one of the world’s fastest-growing megacities.</p>