Greening of the Earth is observed during the past several decades and both climatic and non-climatic factors drive this process. However, the greening spatio-temporal patterns and the role of human activities such as agricultural intensification in hyper-arid regions remains unclear. This study aimed to (i) reveal the greening pattern in southern Xinjiang using satellite estimations of Normalized Difference Vegetation Index (NDVI) and Leaf Area Index (LAI) data during 1982-2019, and (ii) examine the impacts of human activities in terms of land use land cover (LULC) data. Our multi-decadal analysis is ideal to reveal the long-term trends, and favor a better understanding of the anthropogenic effects in this hyper-arid and endorheic region. The results showed that vegetation as a whole increased significantly in southern Xinjiang and the greening rate of cropland was much higher than the other LULC types. Significant greening was found over >90% of cropland, while insignificant changes and browning trends were found over nearly half the area of the other LULC. The proportion of greening areas was more than 80% within 1km from human14 dominated areas while the proportion decreased to 40% with distances more than 15 km. The spatial heterogeneity of the greening indicated that, despite widely reported beneficial effects of warmer and wetter climate for a general greening trend, human activities could be the dominant factor modulating the greening rates disproportionately over different LULC in arid and hyper-arid areas.