The variation in vegetation within semi-arid regions is recognized to exert vital impacts on regional ecological security; nevertheless, the contributions of climate change and human activities dominated by large-scale inter-basin water diversion projects to vegetation dynamics have not been quantitatively separated to date. Taking the water-receiving area of the Tao River Water Diversion Project as the research domain, multi-source datasets spanning 2000–2023 were adopted, and trend analysis, abrupt change test, partial correlation analysis, residual trend analysis as well as nonlinear Granger causality test were implemented to quantitatively partition the respective contributions of climate change and water diversion-oriented anthropogenic interference to vegetation variations. The derived results are summarized as follows. From 2000 to 2023, significant increasing trends were detected for normalized difference vegetation index (NDVI) and precipitation, whereas non-significant rising and declining tendencies were identified for air temperature and potential evapotranspiration (PET), respectively. Abrupt shifts of NDVI, precipitation and air temperature were statistically verified during 2010–2011, while the abrupt change of PET was postponed until 2021. Spatially, areas with extremely significant NDVI growth were observed to cover nearly the entire study territory, and precipitation increments with varied magnitudes were detected across the whole region. The fluctuation amplitude of NDVI was proven markedly larger than that of climatic variables. Partial correlation calculations indicated that NDVI was predominantly positively correlated with precipitation; significant positive correlations between NDVI and air temperature were found in the southeastern zone, and negative correlations were dominant between NDVI and PET. Prior to water conveyance, NDVI variations were jointly driven by climate change and human activities with a high contribution proportion attributed to anthropogenic factors. After the initiation of water supply, regions with vegetation improvement solely driven by human activities were substantially expanded, accompanied by an elevated average contribution rate of human activities and a corresponding reduction in the contribution from climate change compared with the pre-water-conveyance period. Land use transition analysis was further employed to verify the supporting effects of water diversion engineering-dominated human activities on vegetation restoration, and the nonlinear Granger causality test was conducted to confirm the critical role of precipitation-dominated climatic drivers. Certain insights for ecological benefit assessment focusing on water diversion infrastructures in semi-arid zones are provided by this research.