<p>Rapid urbanization is changing the land use and land cover in many parts of the world, dramatically increasing the land surface temperature (LST). This work highlights the seasonally varying influence of various surface biophysical characteristics on the LST of Bhubaneswar, a rapidly urbanizing tropical city in eastern India, from 2003 to 2024. Since the city is mainly dominated by urban land covers, with abundant vegetation and waterbodies scattered throughout, normalized difference indices of built-up, vegetation, water, moisture, bareness, impervious surfaces, and surface albedo were considered for the analysis. Urban indices were found to be most correlated with LST during winter and pre-monsoon. In contrast, water and bareness indices had the highest correlation during the post-monsoon seasons. Overall, averaged LSTs in winter and pre-monsoon were observed to have decreased, while post-monsoon LSTs have increased over the study area. Five statistical and machine learning models (multiple linear, stepwise, random forest, support vector, and neural network) were also used to examine the impact of changing land surface characteristics on LST. The random forest regression model best estimated seasonal variability of LST, as assessed from the low root mean squared error and high R² compared to the others. Inferences from this study would help in better understanding the seasonal variability of LST in response to various urban forces in tropical cities under rapid urbanization.</p>