Understanding the drivers of species distributions in human-dominated landscapes is crucial for proposing sound conservation strategies. Primates are the most studied terrestrial vertebrate taxa, yet still their response to forest loss and fragmentation widely varies among species. In this paper, we assessed the relative influence of local vs landscape features on occupancy of two primate species – the black-fronted titi monkey and the black-pencilled marmoset, in a Brazilian fragmented region. We created detection histories by performing repeated auditory surveys on 25 native vegetation patches. Then we fitted occupancy models using habitat and GIS-based data as site covariates, and weather conditions as detection covariates. We found that forest-like canopy elements are important for the titi monkey, which is a forest-dependent species. Marmoset occupancy was also related to local elements, but in a lesser extent. In addition, we found that ignoring detectability in playback call surveys created a 20% difference in occupancy estimates for the marmoset. We conclude that drivers of primate occupancy at the studied landscape rely mainly on local key habitat elements, so that on-ground conservation actions should not focus on habitat amount alone. Furthermore we reiterate that primate researchers should explicitly account for imperfect detection to avoid substantial detectability bias.