Minimum dataset.

The karst region in southwestern China is particularly prominent and has become a core issue constraining ecological environment restoration and sustainable development in this area. This study utilized long-term remote sensing data to reveal the spatial pattern evolution characteristics of rocky desertification in the region in 2000, 2010, and 2020. Meanwhile, it analyzed the dynamic trend of vegetation coverage recovery in the area from 2000 to 2020, as well as the analysis of related factors. The results showed that the spatial distribution of the Normalized Difference Vegetation Index (NDVI) remained highly clustered, though the clustering gradually weakened over time. When NDVI exceeded 0.6, the probability of rocky desertification reversal increased. Currently, a core contradiction of “quantity increases but quality stagnates” exists in regional vegetation cover, characterized by a continuous rise in NDVI mean values coexisting with reduced spatial clustering. This phenomenon reflects the evolution of vegetation patterns under the combined effects of ecological engineering interventions, adjustments in human-land relationships, and constraints of karst landforms. Through factor analysis, slope and humidity were identified as key factors influencing vegetation restoration. The findings provide an important theoretical foundation and practical reference for targeted rocky desertification management, optimization of ecological restoration projects, and coordinated human-land development in karst regions.