Low detectability of small nocturnal carnivores and biases associated to different census methods hamper the interpretation and reliability of the results of censuses and habitat studies of many cryptic and elusive species, especially because of false-negatives and/or lack of negatives. In order to overcome this problem, methodologies based on the use of presence-only data have been used to predict distribution of species. In this paper, we used presence data of two abundant nocturnal carnivores to test for segregation in their habitat. We compared niche overlap between the common genet and the stone marten at two different spatial scales, home range scale and landscape scale, through logistic regression analyses using presence-only data from Biscay, an area in which both species are common and widespread. We found great niche overlap at both spatial scales, but in spite of it logistic regression analyses found statistically significant differences in the predictor values of some variables. Habitat of genets and stone marten was differentiated by areas with dense vegetation that were of importance for genets, and open areas that were characteristic of stone marten habitat. We suggest that competition between the two species causes the observed segregation.