Abstract: Under the dual pressures of rapid socio-economic development and global climate warming, the impacts of anthropogenic activities and climatic variability on basin-scale hydrological cycles have become increasingly evident. Among the most visible outcomes is the sustained decline in river discharge, particularly in arid and semi-arid regions where water resources are naturally limited. In extreme cases, perennial rivers have degraded into intermittent or seasonal streams, intensifying regional water stress and ecological fragility. Understanding the long-term evolution of runoff characteristics, quantifying the driving forces behind these changes, and distinguishing their respective contributions are essential for improving hydrological modeling, enhancing runoff prediction accuracy, and advancing sustainable water resource management. These challenges remain central yet difficult issues in hydrology and environmental science. The Shichuan River, a major tributary of the Wei River, flows through one of the most water-deficient areas of the Guanzhong Plain. Since the 1970s, extensive human interventions—such as large-scale water extraction, reservoir construction, and land-use changes—have significantly disrupted the natural rainfall–runoff relationship within the basin. These changes have led to a pronounced decline in natural runoff generation, particularly in river reaches characterized by high infiltration capacity, where prolonged flow interruptions have become increasingly frequent. Over the past three decades, the river has experienced 16 distinct drying events, totaling 172 days of zero discharge. Such hydrological disruptions pose significant threats to aquatic ecosystem integrity, water supply reliability, and regional socio-economic sustainability, thereby necessitating a systematic investigation into their underlying causes. To quantitatively assess runoff variation and attribute its driving forces in semi-arid to semi-humid basins with substantial channel infiltration, this study adopts a multi-method analytical framework. First, the Mann-Kendall (M-K) trend test and hierarchical clustering were employed to identify temporal discontinuities in hydro-meteorological time series, thereby delineating distinct hydrological periods. Next, a distributed hydrological model was constructed by coupling the physically based MIKE SHE system with the MIKE 11 hydrodynamic module. This integrated framework explicitly incorporates the effects of human activities (e.g., water withdrawals, land-use changes) and channel infiltration processes on runoff generation and concentration dynamics. Based on this model, a high-resolution spatiotemporal attribution analysis was conducted to quantify the relative contributions of climatic variability and anthropogenic disturbances to the observed runoff decline. The proposed methodology was rigorously validated through its application to the Shichuan River Basin. Key findings include: (1) Temporal Variability in Hydrological Series: Between 1972 and 2020, a statistically significant increasing trend in annual runoff was observed at the upstream Liulin Station, while a decreasing trend was recorded at the basin outlet. No abrupt change was detected in the precipitation series; however, the evaporation series exhibited a marked shift in 1979. Runoff records at Liulin Station showed significant change points in 1976 and 2009, likely attributable to upstream reservoir regulation and climate variability. Based on these breakpoints, the study period was divided into three phases: the Baseline Phase (1972–1979), Change Phase I (1980–2009), and Change Phase II (2010–2020). (2) Attribution of Runoff Changes: Compared with the baseline, climate variability—primarily changes in precipitation and evaporation—was the dominant driver of runoff increases in the reach above Taoqupo Reservoir during Change Phases I and II, contributing 86.6% and 72.1% of the observed changes, respectively. In contrast, human activities, particularly intensive water withdrawals for agricultural and industrial purposes, were the primary contributors to runoff decline in the Taoqupo Reservoir–Fuping and Fuping–basin outlet segments. The contribution of anthropogenic water extraction alone ranged from 146.4% to 183.1%, underscoring its dominant role in exacerbating flow depletion. (3) Channel Infiltration Trends: Across all three periods, riverbed infiltration showed a consistent upward trend, with the Taoqupo Reservoir–Fuping segment exhibiting the highest annual average infiltration per unit river length.