The solar radiation incident in a mountainous area with a complex terrain has a strong spatial heterogeneity due to the variations in slope orientation (self‐shading) and shadows cast by surrounding topography agents (topographic shading). Although slope self‐shading has been well studied and considered in most land surface and hydrological models, topographic shading is usually ignored, and its influence on the thermal and hydrological processes in a cold mountainous area remains unclear. In this study, a topographic solar radiation algorithm with consideration for both slope self‐shading and topographic shadows has been implemented and incorporated into a distributed hydrological model with physically based descriptions for the energy balance. A promising model performance was achieved according to a vigorous evaluation. In a control model without considering the topographic shadows, the simulated solar radiation incident in the study area was about 14.3 W/m2 higher on average, which in turn led to a higher simulated annual mean ground temperature at 4 m (by 0.41 °C) and evapotranspiration (by 16.1 mm/a), and a smaller permafrost extent (reduced by about 8%), as well as smaller maximal snow depth and shorter snow duration. Although the simulation was not significantly improved for discharge hydrograph in the base model, higher river runoff peaks and an increased runoff depth were obtained. In areas with a rugged terrain and deep valleys, the influences of topographic shadows would even be stronger in reality than the presented results, which cannot be ignored in the simulation of the thermal and hydrological processes, especially in a refined model.