Land use land cover (LULC) significantly influences the chemistry, biology, and socioeconomics of a watershed or basin. Additionally, land surface temperature (LST) is a critical factor in understanding climate change. Therefore, analyzing changes in LULC and their drivers, along with the associated shifts in LST, is essential for effective planning and policy development in environmental conservation. This study examines decadal changes in LULC and their impacts on LST in the Kosi River Basin (KRB) in Uttarakhand, using satellite data from the Landsat series. The Kosi River, a spring-fed river, is significant in Uttarakhand, experiencing LULC changes due to factors such as forest fires, outmigration, alterations in agricultural practices, and urban development. The research employs a standard methodology for LULC detection, utilizing NDVI and LST retrieval through single channel (SC) and split window (SW) algorithms for the years 1999, 2009, and 2019. Six categories of LULC were defined: agricultural land, barren land (bare soil and sediment), built-up areas, forests (dense/open), sparse vegetation (grazing), and water bodies. Over the two decades studied, there was a slight decline in forest areas, agricultural land, and water bodies. Conversely, barren land, built-up areas, and sparse vegetation increased. LST data from the Landsat ETM for 1999 shows a temperature range from 11.32 °C to 33.13 °C. In 2009, data from Landsat TM indicates a range from 10.43 °C to 35.64 °C. By 2019, Landsat OLI data shows a range from 2.89 °C to 31.54 °C. The LULC and LST analysis suggest that the retrieval of LST using the SC and SW algorithms, as described in the methodology, reasonably simulates LST in relation to the corresponding LULC for 1999 and 2019, with some exceptions. Overall, there is a marginal difference in LST from 1999 to 2019, with slight variations across different LULC classes. Forested areas consistently exhibited the lowest LST, while agricultural land had the highest LST in 1999, and barren land showed higher LST values in both 2009 and 2019. The mean LST across these years indicated consistently higher temperatures in barren areas associated with other land cover classes. This study further investigates the relationship between LST, NDVI, and NDBI to understand the effects of LULC changes on LST. LST and NDVI exhibited a negative relationship, with R2 values of 0.27 in 1999, 0.21 in 2009, and 0.13 in 2019. In contrast, NDBI and LST exhibit a positive relationship, with R2 values of 0.58 in 1999, 0.48 in 2009, and 0.29 in 2019. The present investigation based on large language models (LLMs) in future land planning has the potential to alter our approach to urban growth, resource management, environmental conservation, and ecosystems resilience. Planners may make informed decisions that combine human needs, environmental sustainability, and economic growth by taking advantage of LLMs’ predictive and analytical skills. This study highlights the significance of geospatial technologies and integrated models in evaluating both historical and current data. LLMs can predict future trends such as population growth, land use changes, and climate impacts, thereby informing sustainable land-planning decisions and efficient management.