Landscape genetics provides a framework for pinpointing environmental features that determine the important exchange of migrants among populations. These studies usually test the significance of environmental variables on gene flow, yet ignore one fundamental driver of genetic variation in small populations, effective population size, N _e . We combined both approaches in evaluating genetic connectivity of a threatened ungulate, woodland caribou. We used least-cost paths to calculate matrices of resistance distance for landscape variables (preferred habitat, anthropogenic features and predation risk) and population-pairwise harmonic means of N _e , and correlated them with genetic distances, F _ST and D _c . Results showed that spatial configuration of preferred habitat and N _e were the two best predictors of genetic relationships. Additionally, controlling for the effect of N _e increased the strength of correlations of environmental variables with genetic distance, highlighting the significant underlying effect of N _e in modulating genetic drift and perceived spatial connectivity. We therefore have provided empirical support to emphasize preventing increased habitat loss and promoting population growth to ensure metapopulation viability.