Over the last six decades, land-cover patterns in Europe have dramatically changed, and major future changes are expected. Land-cover changes affect landscape functions. Therefore, methods are needed to include the temporal dimension into landscape classification. By combining recent satellite data with historic information on land cover from 1955, and the application of k-means cluster analysis, we developed an approach to identify types of land-cover patterns and dynamics (TLPDs) at the rural district scale. Our study area was the Lahn-Dill Highlands, a marginal German landscape with a total of 192 rural districts. We identified six TLPDs that showed a general trend of abandonment, but revealed remarkable differences in current land-cover patterns and the directions of land-cover change. The TLPDs showed notable differences in physical attributes: In the eastern part of the area, where elevation, the proportion of steep slopes, and dry soils are low, land cover remained relatively stable. Slight to dramatic changes occurred, in contrast, in the remaining districts with comparatively unfavourable conditions for cultivation. The spatially differentiated information on areas with contrasting land-cover dynamics within a region may be useful to develop effective concepts for future land management.