Climate and atmospheric CO2 concentration have changed significantly in the mountainous region of the Haihe River basin (in northern China) over the past five decades. In this study, a process-based terrestrial model, version 4 of the Community Land Model (CLM4), was used to quantify the spatiotemporal changes in runoff across the region, driven by the varying climate factors and CO2 concentration. Overall, our simulations suggest that climate-induced runoff in this region display a decreasing trend since 1960. Changes in precipitation, solar radiation, air temperature, and wind speed account for 56%, −14%, 13%, and −5% of the overall decrease in annual runoff, respectively, although their relative contributions vary across the study area. The rise in atmospheric CO2 concentration was determined to have a limited impact on runoff. A significant decrease in runoff in the southern and northeastern portions of the region is primarily attributed to decreasing precipitation, whereas decreasing solar radiation and increasing air temperature were the primary causes of a slight runoff increase in the northern portion. Our results also suggest that the magnitude of the decreasing trend may be greatly underestimated if the dynamic interactions of vegetation phenology and the environmental factors are not considered in the modelling, indicating the importance of including vegetation dynamics in the prediction of runoff trends in this region.