Understanding what factors influence species occupancy in human-modified landscapes is a central theme in ecology. We examined scale-dependent habitat relationships and site occupancy in reptiles across three topographically different study areas in south-eastern Australia. We collected presence–absence data on reptiles from 443 sites associated with three long-term biodiversity monitoring programs, on four to seven occasions, between 2001 and 2013. We characterised sites by the following four variable domains: 1) field design, 2) topography, 3) local-scale vegetation attributes and 4) landscape-scale vegetation cover. We constructed occupancy models for 14 species and used an information-theoretic approach to compare multiple alternative hypotheses to explain occupancy within and between study areas. We modelled detection probability and used the model with the lowest AIC in subsequent analyses. We then modelled occupancy probability against all subsets of the variable groups (field design, topography, local- and landscape-scale vegetation), as well as a model that held occupancy constant (null model). We found that local-scale vegetation attributes were important for explaining site occupancy in 12/19 possible models, although, in several cases model fit was improved by the addition of topographic variables or native vegetation cover in the surrounding landscape. Occupancy models for widespread species were broadly congruent across study areas. We demonstrate that topographic variables are important for explaining reptile occupancy in hilly landscapes, and local- and landscape-scale variables are important for explaining reptile occupancy in flat or gently undulating landscapes. Management actions that improve habitat complexity at a site-level, and encompass entire topographic gradients, will have greater benefit to woodland reptiles than simply increasing vegetation cover in the surrounding landscape.