<p>The COVID-19 pandemic has exacerbated the number of individuals who have reverted to poverty globally, underscoring the significant influence of sudden external disturbances on the global poverty reduction process. Nevertheless, the risk of reversion to poverty has often been overlooked in regions grappling with persistent internal challenges and insufficient capacity for sustainable poverty reduction. Consequently, this study utilized Lushui City, a representative alpine-gorge region in China, as a case study. It selected three pivotal time points in China's poverty reduction timeline (2000, 2010, and 2019) and integrated multi-source geospatial data to establish a regional sustainable poverty reduction capacity (SPRC) evaluation index system. Subsequently, the study employed a backpropagation artificial neural network to assess the SPRC of each village in Lushui City, thereby revealing its spatiotemporal evolution characteristics. By analyzing the lag characteristics of SPRC across various dimensions, the study further categorized the lag types of different villages. Finally, the random forest and multi-scale geographically weighted regression models were utilized to investigate the dominant factors influencing SPRC in Lushui City across different stages and the spatial heterogeneity of their impacts. The findings indicated that: (1) Throughout the study period, Lushui City encountered substantial challenges in achieving sustainable poverty reduction, although implementing poverty reduction initiatives had bolstered its economic sustainable poverty reduction capacity and social sustainable poverty reduction capacity. (2) Some villages in Lushui City exhibited multidimensional lag in their SPRC, necessitating comprehensive measures to achieve regional sustainable poverty reduction. (3) In 2000 and 2010, the primary factors influencing SPRC in Lushui City were environmental (soil erosion amount and precipitation) and social (distance to the nature reserve). However, with the advancement of regional poverty reduction efforts, the main influencing factors in 2019 shifted to economic (land-use mixed degree, density of tourist spots) and social (village agglomeration degree), thereby mitigating the adverse effects of environmental factors on regional SPRC. (4) Most environmental and economic factors exerted a nearly universal impact on SPRC. In contrast, social factors (distance to the main road) exhibited considerable spatial heterogeneity and a limited impact range, with their influence diminishing rapidly beyond a certain threshold. Therefore, enhancing the relevance of social assistance measures and broadening their implementation scope was essential. The recommendations derived from these findings help to provide successful experiences from China for poverty reduction in other alpine-gorge regions.</p>