Drought identification is crucial for various environmental and ecological considerations. This study observed the spatial and temporal variations of drought based on the satellite-derived Vegetation Condition Index (VCI) across different time scales—i.e., winter, pre-monsoon, and annual—compared with remotely sensed Land Surface Temperature (LST) and precipitation. MODIS NDVI products, LST, and meteorological station data for rainfall were used. VCI was employed to classify drought conditions. Pearson correlation analysis was conducted between VCI and both LST and precipitation. The results show that severe drought was detected in 2001 during the winter and pre-monsoon seasons, and also in 2006 during the pre-monsoon season. No severe drought was observed on the annual time scale, although moderate droughts occurred in several years. The VCI trends increased at rates of 1.29 yr⁻¹, 1.52 yr⁻¹, and 1.72 yr⁻¹ for the annual, winter, and pre-monsoon periods, respectively. Conversely, LST decreased at rates of 0.007 yr⁻¹, 0.044 yr⁻¹, and 0.028 yr⁻¹ during the same periods. The increased VCI and decreased LST indicate a declining trend in drought severity on a yearly scale. The positively increasing Anomaly of VCI (AVCI) suggests improved vegetation growth under moist soil conditions. VCI was also positively correlated with precipitation on the annual and pre-monsoon time scales, but negatively correlated during the winter season. The correlation between LST and VCI was negative across all time scales and was more significant during the pre-monsoon and winter seasons. This study contributes to understanding vegetation-based drought and its relationship with temperature and precipitation in the Koshi Basin, Nepal.