Large old trees (LOTs) play critical ecological, historical, and cultural roles but are experiencing rapid population declines globally. Climate change, particularly rising greenhouse gas concentrations, further threatens their survival. Understanding the determinants of LOT distribution is essential for their conservation, especially in the upper and middle reaches of the Tarim Basin. This study presents the first large-scale survey of LOTs in arid land of northwestern China, combines field expeditions with Generalized Linear Models (GLMs) and Random Forest (RF) analysis, we: (1) documented LOT spatial distribution, revealing 89.33 % of trees were <300 years old, with Populus spp. and Juglans regia as dominants; and (2) identified key drivers of LOT persistence. Our results highlight water availability, population density, soil properties, and climatic conditions as key determinants of LOT longevity and health condition. GLM analysis identified proximity to water systems ((3 = -0.012, p < 0.001) and population density ((3 = -11.704, p < 0.001) as significant factors, while RF results emphasized proximity to water systems, road infrastructure, soil cation exchange capacity, annual precipitation and temperature seasonality. These findings underscore the urgent need for government action to address groundwater loss and promote public participation in LOT conservation in arid regions.