Abstract: The Yellow River Source Region is located in the core area of Sanjiangyuan National Park on the northeastern Tibetan Plateau, and it is highly sensitive to climate change and ecohydrological regulation. In this alpine headwater basin, river morphology serves as an important indicator that captures the coupling between hydrological processes and ecosystem dynamics. Among various morphological indices, river width is a key parameter for describing hydrodynamic conditions and reflecting how river systems respond to environmental change. However, a sufficient understanding of the multi-scale linkage mechanisms between river width variations and environmental drivers in the Yellow River Source Region is still lacking. In particular, conventional interannual analyses often mask the short-term fluctuations of hydrological and meteorological elements at monthly and quarterly scales. In this study, 20 representative river cross-sections were selected across three geomorphic zones: the headwater valley, the mountain gorge section, and the basin meandering section of the main stem. Using the JRC Global Surface Water dataset and an improved RivWidthCloud workflow implemented on Google Earth Engine, a time series of river width at annual, quarterly, and monthly scales was constructed for the period 2001—2020. To reduce the discontinuity of monthly data caused by cloud cover and satellite revisit cycle limitations, Sentinel-2 imagery was introduced to fill missing monthly observations, and the resulting monthly river width retrievals were calibrated against Landsat data. Meanwhile, 21 environmental variables covering climatic, hydrological, vegetation, and human-related conditions were integrated. Pearson correlation analysis was applied at the interannual scale, while interaction analysis based on CEEMDAN and monthly lag analysis were applied at the quarterly and monthly scales. The short-term relationships between environmental factors and river width variations, as well as their temporal characteristics, were systematically explored. The results showed that the river width dataset extracted in this study agrees well with the Global River Widths from Landsat dataset, with an R² of 0.84 and a mean absolute percentage error of 11.4%. This indicates that the proposed framework is applicable to complex alpine river systems. At the interannual scale, over 75% of the cross-sections showed a positive correlation between river width and precipitation, and the vast majority of cross-sections exhibited a negative correlation between river width and surface temperature. This suggests that, under the background of warming and wetting, the long-term adjustment of river channels is closely related to regional water and heat conditions. At the monthly scale, fluctuations in river width were more strongly associated with short-term moisture-related processes, especially precipitation, soil moisture, and evapotranspiration. According to the lag effect analysis, the influence of these variables on river width generally had a lead time of 0—1 month, indicating a relatively concentrated short-term response. In contrast, air pressure and wind speed showed weaker consistency in their lag patterns, and their influence on river width was more indirect through meteorological regulation rather than direct hydrological driving. At the quarterly scale, the differences in the effects of individual environmental factors diminished compared with the monthly scale, and river width variations exhibited a more integrated response pattern. This indicates that as the temporal window expands, the directional differences in short-term responses are progressively smoothed out. Overall, river width variations in the Yellow River Source Region exhibit significant timescale dependence, meaning that the dominant environmental drivers differ markedly across temporal scales. These findings deepen the understanding of multi-scale river evolution processes in alpine headwater basins and provide a scientific basis for hierarchical monitoring and adaptive ecological management of Sanjiangyuan National Park.