Urban and environmental studies have long been interested in understanding the connection between land surface temperature (LST) and land use/land cover (LULC). Artificial intelligence models, regression and decision trees have been used in previous research to investigate this. This study examined the links between LULC and LST using the explainable artificial intelligence (XAI) technique to get over the drawbacks of earlier models. We created the LST prediction model in Bangalore by combining the XGBoost and SHAP models, and we calculated the effects of LST reduction following LULC modifications. The results demonstrated that using LULC, topographic and landscape data not beyond the buffer radius as independent factors increased the forecast accuracy of LST. In particular, the presence of nearby vegetation and built-up areas was determined to be the most significant contributing factor to the explanation of LST. This study indicated that the LST will decrease by about once the LULC changed from the freeway to green areas. Our study's conclusions can be used to evaluate and track how urban planning, and initiatives affect the thermal environment. Additionally, this analysis may help priorities various governmental initiatives aimed at enhancing the thermal environment.