Abstract: Desertification risk assessment in arid regions was frequently hindered by the coarse spatial resolution of available soil datasets and the inherent limitations of traditional linear evaluation methodologies. These conventional approaches often failed to elucidate the complex dynamic and static coupling mechanisms driving ecosystem evolution in highly fragmented habitats. Consequently, accurately identifying specific risk sources and micro-geomorphic variations remained a significant challenge for ecological management and sustainable development. To address these scientific gaps, this study aimed to construct a high-precision, comprehensive desertification risk assessment system tailored for extreme arid zones, thereby revealing the spatiotemporal evolution patterns of ecosystem risks. The Tarim Populus euphratica National Nature Reserve, a typical ecologically fragile region, was selected as the study area. A novel analytical framework was developed and implemented. Initially, a machine learning approach utilizing the Random Forest (RF) algorithm was employed to spatially downscale soil texture data from a 1 km resolution to 30 m. This step integrated high-resolution remote sensing predictors with empirical field sampling data to enhance baseline accuracy. Subsequently, the study formulated a Comprehensive Desertification Risk Level (CDRL) through a hierarchical zoning matrix. This matrix effectively coupled the dynamic wind erosion driving forces, quantified by the Revised Wind Erosion Equation (RWEQ), with the static ecosystem sensitivity, characterized by the Desertification Sensitivity Index (DSI). Furthermore, the Theil-Sen Median trend analysis combined with the Mann-Kendall test, alongside the Optimal Multivariable Stratified Geodetector (OMGD), was applied to quantitatively decode the spatiotemporal evolution characteristics and the multi-factor nonlinear driving mechanisms of regional desertification risks over the period from 2000 to 2024.The comprehensive analysis yielded several critical findings. First, the spatial downscaling process demonstrated robust performance, achieving R² values ranging from 0.79 to 0.82, alongside a Kling-Gupta efficiency (KGE) between 0.64 and 0.70, when validated against in-situ measurements. This high-resolution mapping effectively captured the spatial heterogeneity of micro-geomorphic features that were previously obscured. Second, in terms of temporal evolution, the area proportion of each desertification risk grade showed significant differentiation from 2000 to 2024. The proportion of extremely low-risk areas increased steadily from 22.16% to 26.99%, indicating the continuous improvement of the ecological environment in the core reserve. The proportion of extremely high-risk areas, despite inter-annual fluctuations, decreased slowly from 31.81% to 28.75% on the whole. The proportion of low-risk areas showed a shrinking trend, while the proportions of medium-risk and high-risk areas fluctuated within a relatively stable range, without presenting a significant increasing or decreasing trend. Third, temporal trend analyses revealed that the overall ecological status of the reserve maintained a resilient baseline, with 73.64% of the area exhibiting no significant changes throughout the study period. However, a pronounced spatial divergence emerged in the evolving areas: 12.82% of the area, mainly located on the north bank, experienced significant risk reduction, while 13.54% of the area on the southern edge faced ongoing degradation, maintaining a delicate dynamic balance. Fourth, the driving force detection indicated that soil properties functioned as the dominant baseline factors, with their explanatory power (q values) consistently exceeding 0.91. Notably, the explanatory power of the vegetation quality index increased significantly from 0.69 in 2000 to 0.82 in 2024, confirming the growing positive impact of anthropogenic ecological water conveyance projects.In conclusion, the proposed evaluation framework successfully facilitated a fine-scale, mechanism-based diagnosis of desertification risk, effectively overcoming traditional coarse-scale shortcomings. The findings indicated that the reserve was experiencing a phase of dynamic balance between ecological restoration and natural degradation. Based on these mechanistic insights, it was strongly recommended that future environmental management adopt a differentiated spatial strategy. This approach should focus on maintaining the ecological connectivity of the core river corridor while implementing targeted engineering governance for the 13.54% of degrading areas on the south bank.