Climate change and human activities are both major drivers of vegetation dynamics. Distinguishing their relative contributions to vegetation change is crucial for formulating reasonable vegetation restoration strategies. Based on remote sensing data, this study employed trend analysis to reveal the spatiotemporal characteristics of normalized difference vegetation index (NDVI) changes in the arid region of northwestern China (ARNWC) from 2000 to 2020. Furthermore, a random forest model was utilized to reveal NDVI changes driven by human activities, followed by an aridity gradient analysis. The results showed that vegetation restoration was widely observed across ARNWC, with 59.6% of the area experiencing a significant increase in NDVI, particularly in the eastern region. However, there were also declines in NDVI, especially in the western region. The majority of climate-induced vegetation changes were positive, with 73.8% of the area exhibiting a significant enhancement in NDVI. In contrast, 56.4% of human-driven vegetation changes were negative. Moreover, human-induced vegetation restoration intensified with increasing aridity, and human activities significantly contributed to forest restoration in areas with the highest aridity. These findings provided critical insights for vegetation restoration in arid regions.