Dispersal is often viewed as a process on which the landscape has little effect. This is particularly apparent in populations’ genetic and ecological studies, where isolation by distance is generally tested using a Euclidean distance between populations. However, landscapes can be richly textured mosaics of patches, associated with different qualities (e.g. different costs crossing patches) and different structures (shape, size and arrangement). An important challenge, therefore, is to determine if accounting for this additional complexity enriches our understanding of the dispersal processes. In this study, we quantify the effect of landscape structure on dispersal distances between 15 populations of the greater white-toothed shrew (Crocidura russula) in a highly fragmented landscape in Switzerland. We use a spatially explicit individual-based model to simulate C. russula dispersal. This model is designed to account for movement behavior in heterogeneous landscapes. We explore the relationship between simulation results and genetic differentiation between actual subpopulations. Finally, we test if simulated dispersal distances are better predictors of genetic differentiation than traditional Euclidean distances. The ecological distances measured by the model show a clear relationship with genetic differentiation between C. russula subpopulations. This relationship is stronger than the one obtained by the usual Euclidean distance.