Vegetation dynamics and land use information are significant for a better understanding of the ecological consequences of multiple mining activities. However, the high spatial heterogeneity of mine sites and diverse disturbance and recovery pathways make it a challenge to understand the dynamics of multiple mine sites over large areas. Here, we proposed a general framework for continuous monitoring of land use and vegetation dynamics in multiple mine sites and applied it to Pingxiang, China. First, annual land use and land cover (LULC) maps from 2000 to 2019 were generated using a modified Continuous Change Detection and Classification approach (CCDC). Second, the locations and extents of 86 mine sites on different scales were mapped individually and then aggregated into five groups according to the similarity and differences of vegetation change. Vegetation dynamics showed great heterogeneity across sites driven primarily by the spatial-temporal variation in types and intensity of land use activities in and around the mine sites. We found the impact distance was typically 500-700 m in the region, but can be smaller than the potential impact distance in areas with land use activities. The long-term slow recovery of vegetation conditions at some sites indicates that it might be a challenge to improve vegetation conditions naturally in a short time and human-assisted restoration measures may be required. The systematic framework proposed in this study can be used to establish comprehensive and spatially-explicit mine datasets at the regional scale, essential for understanding the dynamics and ecological consequences of multiple mining activities and coordinated management and restoration of heterogeneous mine sites.