Abstract: In order to deeply analyze the mechanisms of different driving or influencing factors on runoff processes, based on the runoff fractionation method (Delayed-flow index, DFI) and group-based multi-trajectory modeling (GBMTM), in conjunction with a random forest and SHAP hybrid model, this study systematically analyzed the distinct impacts of meteorological factors, underlying surface conditions, and anthropogenic activities on various runoff components.The research additionally examined the impact of regional geomorphic features on the development of river runoff elements.The findings indicate that four runoff components can be separated from the measured runoff series, among which the Short-term component is predominantly influenced by basin-wide meteorological factors, the Intermediate-term component is mainly restricted by underlying surface conditions, while the Long-term and Baseline components are more significantly affected by human activities on the surface, displaying certain hysteresis and regional characteristics.Regional analysis indicates that natural processes dominate the northern foothills of the Qinling Mountains, human activities are prevalent in the Guanzhong Plain, and the Loess Hilly Region is affected by a blend of meteorological elements and anthropogenic surface activities.These findings offer scientific evidence for clarifying the mechanisms of runoff formation and their reactions to anthropogenic disturbances at the component level.They additionally provide a technical reference for further investigation into watershed water resource regulation and the evaluation of human activity impacts.