Land surface temperature (LST) is a key parameter in land–atmosphere interactions. The recently released Moderate Resolution Imaging Spectroradiometers (MODIS) LST Version 5 products have provided good tools to evaluate water and energy budget modelling for river basins. In this study, a distributed biosphere hydrological model (WEB-DHM; so-called water and energy budget-based distributed hydrological model) that couples a biosphere scheme (SiB2) with a geomorphology-based hydrological model (GBHM), is applied to the upper Tone River Basin where flux observations are not available. The model facilitates a better understanding of the water and energy cycles in this region. After being calibrated with discharge data, WEB-DHM is assessed against observed streamflows at four major gauges and MODIS LST. Results show that long-term streamflows including annual largest floods are well reproduced. As well, both daytime and nighttime LSTs simulated by WEB-DHM agree well with MODIS observations for both basin-averaged values and spatial patterns. The validated model is then used to analyze water and energy cycles of the upper Tone River Basin. It was found that from May to October, with relatively large leaf area index (LAI) values, the simulated daily maximum LST is close to soil surface temperature (Tg) since Tg is much greater than canopy temperature (Tc) in their peak values; while the daily minimum LST appears similar to Tc. For other months with relatively small LAI values, the diurnal cycles of LST closely follow Tg.