To conserve areas and species threatened by immediate landscape change requires that we make planning decisions for large areas in the absence of adequate data. Here we study the utility of broad-scale landscape metrics as predictors of species occurrence, especially for remote areas where there is a need to make the most of limited spatial and biological data. Bonobos (Pan paniscus) are endangered great apes endemic to lowland forests of the Democratic Republic of Congo. They are threatened by bushmeat hunting that is exacerbated by habitat fragmentation through slash-and-burn agriculture and timber harvest. We developed four landscape metrics —edge density (ED), COHESION, CONTAGION, and class area (CA)— that may serve as surrogates for measuring accessibility of areas to hunting in order to predict relative bonobo-habitat suitability. We calculated the metrics for the Maringa-Lopori-Wamba (MLW) landscape and evaluated them for utility in predicting bonobo-nest occupancy based on 2009 field data. Cross-validations showed that all four metrics performed similarly. However, forest ED was arguably the best predictor, with an overall classification accuracy of 72.1% in which 85% of known nest blocks (N0124) were classified correctly. We demonstrated that for a relatively intact landscape and a mobile forestdwelling species that is fairly tolerant of forest openings, forest fragmentation can still be an important predictor of species occurrence. We suggest that ED can be helpful when mapping bonobo habitat in MLW and can aid landscape-planning and conservation efforts. Our approach may be applied to other edge-sensitive species, especially where high-resolution data are deficient.