Vegetation often exhibits complex dynamic trajectories under the influence of multiple external factors. In arid regions, fragile ecosystems and high climate sensitivity make vegetation more responsive to environmental fluctuations, resulting in inadequately captured dynamics by a single linear trend analysis. Therefore, this study employed a multi-trajectory model to characterize vegetation dynamics across China's arid region from 2000 to 2024, followed by a random forest analysis to identify the natural and anthropogenic drivers of different trajectory types. The results showed that: (1) Vegetation in the arid region exhibited an overall greening trend, with approximately 74 % of the area showing vegetation improvement, while only about 7 % experienced a decline. (2) Abrupt-change trajectories were the dominant vegetation evolution pattern, occupying 49 % of the region. Linear and curvilinear trajectories accounted for 17 % and 15 %, respectively, and 19 % of the area showed no significant trend. (3) The Random Forest analysis revealed that soil properties and increased annual precipitation were the most influential factors promoting positive trajectories, while urbanization and increased droughts negatively affected vegetation. This study provides a deeper understanding of vegetation dynamics and their drivers in arid regions and offers scientific support for regional ecological management and sustainable development.