The degradation of permafrost on the Tibetan Plateau (TP) exacerbates the expansion of retrogressive thaw slumps (RTSs), posing a threat to the stability of infrastructure. Based on time-series Landsat images, we used a combination of machine learning automatic classification and manual delineation to identify 3537 RTSs in the permafrost region of TP from 1986 to 2020, the data also provided detailed information on RTSs area and annual disturbance time.