Karst rocky desertification (KRD) is a typical ecological issue in southwestern China. Comprehensively revealing the KRD succession processes is crucial for implementing effective restoration efforts. In this study, we proposed a generalized framework to improve the monitoring of KRD at regional scale using high-frequency Landsat time series. The framework involved the development of the Karst Rocky Desertification Criterion Index (KRDCI) and multi-temporal scales trends analysis. We illustrated the use of this framework in Puding County southeast China, by (1) exploring the KRD succession processes at three time scales: General Trend (GT), Inter-annual Cumulative Trends (IACT), and Intra-annual Variability Trends (IAVT) from 2000 to 2021 and; (2) modeling and quantifying the driving factors. Results showed that: (1) KRDCI is a highly effective indicator, with overall accuracy and Kappa coefficients above 0.9; (2) The GT analysis showed a positive improving trend with 59.3 % areas improved and 40.7 % degraded, only 1.4 % areas were persistent significant degraded. Indeed, the IACT analysis illustrated that degraded areas decreased from 64.8 % to 45.7 % (2000-2003) and 50.7 % to 40.7 % (2013-2021), while increased from 45.7 % to 50.7 % during 2003-2013. In contrast, the IAVT analysis indicated that improved areas ranged from 22.9 % to 40.5 % and degraded areas ranged from 59.5 % to 77.1 % over 22 years. In particular, the persistent significant degraded areas remained at 25.0 % for almost all the years. (4) Random Forest demonstrated that environmental factors and land use change are both major influences, and extreme climate events have long-lasting legacy effects on KRD succession. These findings indicated that the vulnerability and low resilience of ecosystem in KRD regions still remained after the implementation of the restoration programs. Although our framework was specific to southeast China, we believed that it would be valuable to monitor and evaluate the effectiveness of restoration efforts in regions characterized by high landscape heterogeneity.