Terrestrial ecosystems are always changing due to natural and anthropogenic factors. Many studies have described the changes through landscape indicators, but landscape stability and processes are considered less in existing studies. Therefore, it is essential to establish effective indicators to detect the landscape changes and stability. In this study, three indicators were proposed to characterize the vegetation landscape stability and processes based on the system fluctuation and stability theory, which were applied to a typical semiarid grassland coal mine region. The indicators were calculated based on the normalized difference vegetation index (NDVI) time-series data griding, which characterized the NDVI central tendency, heterogeneity, and relative fluctuation in each grid through spatio-temporal analysis. The study has confirmed that the dynamic process and stability of each landscape type identified through the three indicators were in good agreement with the on-site conditions. Firstly, the natural grassland was in a macro-stable state generally. Secondly, partial regions associated with open-pit mining regions were identified as severe disturbance regions, with grassland entirely replaced by the bare land, while the grasslands surrounding were identified as mild disturbance regions, with the stability decreasing. Finally, the surface subsidence was characterized by ground surface deformation and fracture, which causes mild disturbance to the grassland landscape. However, the surface collapse, characterized by subsidence pits, has always caused grassland instability. Moreover, the reclaimed grassland is prone to deterioration and lack of stability under severe natural conditions such as drought. We suggest this study could be used to identify the stability and dynamic trend of grassland landscape.