Identification of habitat features that are strongly associated with the occurrence of threatened species is important in terms of predicting impacts of habitat change and identifying key habitats for conservation. In this paper, we apply habitat-based statistical models to predict occupancy patterns of the endangered Southern Bell Frog (Litoria raniformis) across inland New South Wales (Australia). Litoria raniformis previously occupied a wide range of natural and man-made waterbodies across a large geographic range, including floodplain wetlands, oxbow lagoons, irrigation canals and rice bays. Alteration of natural flooding regimes and livestock grazing have affected a large proportion of habitats within the historical range of L. raniformis, but it is not clear how these changes have influenced habitat occupancy patterns. Fifty-two waterbodies were surveyed for presence/absence of L. raniformis in 2001 and 2004. Step-wise logistic regression models were generated to select a subset of variables that could best predict occupancy. The interaction between wetland hydrology and complexity of aquatic vegetation, along with the complexity of fringing vegetation and water temperature were able to delineate between vacant and occupied waterbodies with an accuracy of 90% for vacant habitats and 70% occupied habitats. While this study demonstrated that a range of water body types are occupied by L. raniformis, these habitats do share common hydrological conditions and vegetation characteristics. Altered flooding regimes and reductions in the complexity of aquatic and fringing vegetation are likely to increase the probabilities of localised extinction events of L. raniformis. populations.