Rapid urbanization and agricultural expansion have dramatically intensified land cover change (LCC) in Chengdu City, China, over the past three decades. This study investigates the interaction between LCC and land surface temperature (LST) from 1990 to 2020 using satellite remote sensing data in combination with cloud computing, machine learning, and reference data migration. A consistent time-series of land cover maps at 30 m resolution was generated for four benchmark years (1990, 2000, 2010, and 2020). The maps achieved an average overall accuracy of 91% and F1-scores exceeding 85% across all classes, confirming the robustness of the classification. The results reveal pronounced LCC driven primarily by agricultural expansion and urban growth. LST analysis showed substantial temporal variability, with mean temperatures stable between 1990 and 2000 (26.27 °C), rising sharply by 2010 (31.86 °C), and declining to 23.69 °C by 2020. Zonal statistics demonstrated strong relationships between LC and LST: bare land (32.49 °C) and built-up areas (28.21 °C) exhibited the highest mean LSTs, whereas forests (21.20 °C) and water bodies (20.33 °C) acted as significant cooling elements. Overall, this study underscores the critical role of LCC in shaping Chengdu’s thermal environment. The results provide new insights into the coupled processes of urbanization, agricultural intensification, and climate variability, while demonstrating a robust and scalable framework for LC monitoring to support sustainable urban and environmental planning.