This study aimed at evaluating the potential role of persistent soil seedbanks in restoring grassland and heathland communities in sites that have been afforested or are currently cultivated. To do so, we used the results of 16 case studies and an analysis of the database of Thompson et al. (1997). Generally, there is a decrease in total seed density with increasing age since abandonment or transformation. This decline is much faster for species of calcareous or alluvial grasslands than for heathland and grassheath species. During succession, seeds of target species disappear from the seedbank, owing to seed senescence. In many cases, the seed density of nontarget species increases with time. These species are often ruderal or competitive species, typical of arable land and other disturbed sites, with long-term persistent seeds. Grassland species have a low seed longevity index and low densities in the soil. Moreover, a few years after abandonment, almost all species have disappeared from the soil seedbank. Hence, restoration of species-rich grasslands cannot rely on the seedbank. Where the colonization by target species is hampered because of the isolation of the site, seed introduction is necessary. Heathland species, especially Calluna vulgaris, and grassheath species have a higher seed longevity index and can survive for a long period in the soil. However, the density of their seeds also declines with time and restoration of heathland on sites abandoned or forested for longer than 40 years is also seed limited. If the former heathland was converted to arable land, the decline of heathland species in the seedbank is much faster owing to ploughing. Moreover, high nutrient levels in the soil favour competition from nontarget species after germination. In that case, restoration of heathland or grassheath is expected to be problematic.