In order to understand the capacity of habitats to conserve species, many authors have searched for a species–area relationship (SAR) to evaluate the effect of patch size on species richness in habitat fragments. However, a range of different processes may underlie or obscure this relationship. For woody plant species in forest fragments, as for other taxa, considering forest edges separately in the investigation of SAR is particularly relevant. The objective of our study was to evaluate edge influence on SAR in a fragmented forest landscape in south-western France. To achieve this objective, we considered SAR in two separate spatial portions of the forest fragment: the edge and the forest interior. We also considered SAR for different ecological groups of species based on their shade tolerance and their mode of dispersal, as species can respond differently to habitat reduction depending on their ecological characteristics. In 40 woodlands in an agricultural landscape, we observed the presence/absence of all woody species along parallel walking transects 20-m wide to inventory the total number of species in each woodlot. Vegetation surveys resulted in the identification of 53 woody species, with a total of 23 trees and 30 shrubs, and a number of species per woodlot ranging from 18 to 44. The species richness found in a given woodlot was significantly correlated with its area. When considering the edge and the interior parts of the woodlot separately we found a steeper increase in species richness with area for the part we considered as the edge, compared with the increase found in the interior part of the forest. Our results confirm the contribution of forest edge to forest fragment SAR. Results also underlined the importance of two additional processes that may contribute to SAR: a possible extinction debt of shade tolerant species in forest edge zones and colonisation by light-demanding species in forest interior zones probably due to disturbances. This approach underlines the importance of taking the edge effect into account when studying the influence of patch size on species richness and the dynamic of species richness pattern.