Understanding catchment hydrological response to intensive land use/cover change (LUCC) and climate change provides a basis for taking effective measures for the future. Runoff is a critical indicator of catchment hydrological processes that reflects the combined effects of climate changes and local human activities. In this study, three main tributary sub-catchments underlain by soft sandstone in the Yellow River basin, China, were chosen to attribute runoff variations to climatic change and human activities through improving the Budyko elasticity model. The results suggested that: (1) annual runoff exhibited a significant decreasing trend during the past 30 years (1981–2016, p < 0.01),with an average decline rate of 1.07 mm a−1; (2) the precipitation elasticity of runoff (εP) and that of potential evapotranspiration (εEo) varied from 2.42 to 2.96 and from −1.96 to −1.42, respectively, indicating that runoff is more sensitive to changes in P than those in Eo in the context of climate change; (3) the attribution analysis demonstrated that, on average, vegetation change (mainly anthropogenic vegetation coverage increase) accounted for 92% of the decline in runoff whereas climate change (including precipitation and potential evapotranspiration variations and consequent vegetation change) accounted for the rest 8%. Display omitted •Climatic and anthropogenic impacts on vegetation changes are distinguished•Annual runoff displayed a significant downward trend during 1981–2016•The annual mean NDVI presented an obvious upward trend•Runoff decline is mainly caused by anthropogenic LUCC