The analysis of changes in the catchment vegetation cover as a result of industrial development of its territory, primarily deforestation, was carried out. For processing Landsat images, the spectral space modeling method was used, which provides higher reliability of the classification results compared to traditional methods. The most significant changes in the catchment occurred from the 1930s (the beginning of large-scale logging), until the end of the 1990s. Significantly, by more than 60 %, the area of old-growth forests decreased, their large fragments remained only in nature protected areas, as well as around large lakes and along large rivers, due to the presence of water protection zones. In the period from 2000 to 2018 due to deforestation, the area of productive forests decreased even more. Nevertheless, by 2019 the occurrence of secondary reforestation on significant areas (up to 70 %) is noted, which indicates a sufficiently high regeneration potential of the catchment ecosystems. To restore the spatial structure of the forest cover in the period before the start of the Landsat 5 operation, the method of tracking the trajectories of reforestation successions in the spectral space was used. A comparison with the results of the Global Forest Change 2000—2018 data (loss/gain) showed almost complete coincidence in cuttings (loss), but the method of reforestation trajectories showed many times higher results in forest gain, since it reveals forest regeneration at a much earlier stage. The developed methodology will be applied on the catchments of other tributaries of the White Sea, and the results will be used as input to an environmental-socio-economic cognitive model that predicts the state of ecosystems under climate changes and economic priorities.