Desertification is one of the most critical global environmental challenges, with significant ecological and socioeconomic impacts. Understanding its dynamics and underlying drivers is crucial for developing effective prevention and control strategies. This study provides a comprehensive spatiotemporal analysis of desertification dynamics in the drylands of Central Asia (CA) from 1982 to 2020, employing the Desertification Difference Index (DDI). A non-linear Granger causality framework was utilized to quantify the relative contributions of natural and anthropogenic drivers to desertification across various land cover types. The trend analysis revealed that approximately 14.81% of the drylands in Central Asia experienced desertification during the study period. Pixel-wise fitted random forest Granger causality models identified Snow Water Equivalent (SWE) as a dominant factor influencing dryland changes in the region, particularly in cropland, grassland, shrubland, and bare land. In contrast, temperature emerged as the dominant driver of forest degradation. Overall, 69.1% of desertification in CA was attributed to natural drivers, while anthropogenic drivers accounted for 30.9%. These findings offer pivotal insights into the processes driving desertification in Central Asia's drylands, highlighting the need for region-specific land management and ecological conservation strategies.