Based on the measured runoff data from Gangcha Station in Shaliu River Basin from 1960 to 2022, the characteristics of runoff change of Shaliu River were analyzed using the Mann-Kendall test and cumulative distance level method. Attribution analysis was carried out using elasticity coefficient method based on Budyko hypothesis. Furthermore, the results of the attribution analysis were verified using the Partial Least Squares Structural Equation Modeling (PLS-SEM). The results showed that: (1) the annual runoff of the Shaliu River basin increased significantly from 1960 to 2022, with a warmer and wetter climate trend; a mutation in runoff occurred in 2003, after which the multi-year average runoff increased by 1.152×108m3, with a relative change rate of 47.2%. (2) The runoff in the Shaliu River basin was most sensitive to precipitation changes, followed by changes in underlying surface, and least sensitive to potential evapotranspiration. The contribution rates of precipitation, potential evapotranspiration, and underlying surface changes to runoff were 52.75%, 0.18%, and 47.08%, respectively, indicating that climate change is the main cause of the runoff increase. The PLS-SEM analysis further confirmed that precipitation is the primary climatic factor driving runoff changes, consistent with the attribution analysis results. (3) The influence of underlying surface changes on runoff changes is positive, resulting from the combined effects of land use changes, permafrost degradation and other factors. Future research still needs to quantify the contribution of permafrost degradation to runoff changes. The study results can provide references for understanding runoff changes in the Shaliu River basin, rational water resource development, and the formulation of ecological policies.