Edge effect involves microclimatic variations (light, temperature and humidity) in the forest edge that become unsuitable for organisms adapted to forest interior conditions. We speculate that under similar edge types (shrubland matrix), the pattern in the response to forest edge might be different and more hostile for epiphytes in south-oriented edges. Because different matrix types have different effects on microclimate, we hypothesized that the magnitude of the edge effects will depend on the contrast between the physiognomy of the matrix and the forest fragment. The study was carried out in central Spain and consisted of a beech fragment surrounded by different types of vegetation matrices: pine plantation (south orientation), semi-deciduous oak forest (south orientation) and shrubland (north and south orientations). For each type of edge, we established ten transects perpendicular to the edge, running 100 m into the forest patch. Data were collected from three plots placed at 5, 50 and 100 m from the beginning of each transect. The results revealed that variations in the epiphytic composition mainly showed a replacement of forest interior species with generalist light-demanding species at forest edge. Shrubs-bordered edges showed the strongest influence on epiphytic composition, with high species dissimilarity between edge and forest interior, whereas the edge effects were mitigated when the patch was surrounded by other forests. Besides, in the case of shrub-bordered edge, the dissimilarity of species composition between edge interior was higher in the south than in the north-facing edge. We conclude that the magnitude of the edge effect on epiphytic communities is modulated by the physiognomy of the surrounding matrix and the orientation of the forest edge. In the case of a hard contrast matrix, at least 100 m of an edge buffer are required for the conservation of the species with high humidity requirements. Therefore, one of the main problems affecting the loss and even disappearance of shaded species is the decreasing of the amount of core conditions under the increasing forest fragmentation scenarios. This potentially exacerbates forest fragmentation effects linked to livestock use, due to the physiognomically different surrounding matrix compared to forest patches.