Aim: Andean montane forests are biodiversity hotspots and large carbon stores and they provide numerous ecosystem services. Following land abandonment after centuries of forest clearing for agriculture in the Andes, there is an opportunity for forest recovery. Field-based studies show that forests do not always recover. However, large-scale and long-term knowledge of recovery dynamics of Andean forests remains scarce. This paper analyses tropical montane forest recovery trajectories over a 15-year time frame at the landscape and tropical Andean scale to inform restoration planning. Methods: We first detect “potential recovery” as areas that have experienced a forest transition between 2000 and 2005. Then, we use Landsat time series analysis of the normalized difference water index (NDWI) to classify four “realized recovery” trajectories (“ongoing”, “arrested”, “disrupted” and “no recovery”) based on a sequential pattern of 5-yearly Z-score anomalies for 2005–2020. We compare these results against an analysis of change in tree cover to validate against other datasets. Results: Across the tropical Andes, we detected a potential recovery area of 274 km2 over the period. Despite increases in tree cover, most areas of the Andes remained in early successional states (10–25% tree cover), and NDWI levelled out after 5–10 years. Of all potential forest recovery areas, 22% showed “ongoing recovery”, 61% showed either “disrupted” or “arrested recovery”, and 17% showed “no recovery”. Our method captured forest recovery dynamics in a Peruvian arrested succession context and in landscape-scale tree-planting efforts in Ecuador. Main conclusions: Forest recovery across the Andes is mostly disrupted, arrested or unsuccessful, with consequences for biodiversity recovery and provision of ecosystem services. Low-recovery areas identified in this study might be good candidates for active restoration interventions in this UN Decade on Restoration. Future studies could determine restoration strategies and priorities and suggest management strategies at a local planning scale across key regions in the biodiversity hotspot.