Abstract Species need to disperse at a broad range of spatial scales, the recognition of which has spawned programmes such as Wildlands and WildCountry that aim to restore large-scale connectivity. To achieve connectivity, a first step is to understand how wildlife uses existing remnants. In this study we examine the effect of remnant isolation and condition on the reptile fauna of fragmented mallee habitats in southern Australia. In three replicate landscapes we use pitfall traps to survey reptiles in five landscape elements: Conservation Park, connected, disconnected and isolated fragments, and the agricultural matrix. Reptile species richness, abundance, abundance of snakes, skinks and the 10 most common species had no significant association with landscape elements, excluding the matrix. This was despite a substantial reduction in plant species richness in the fragments, particularly of shrubs. Only seven individual reptiles were captured in the matrix, most on one site with deep sandy soils. The farmland on clay soils appeared to be relatively impermeable for reptiles, although four species could traverse 100 m of cleared sand-dune. The lack of an isolation effect suggests that populations in remnants are persistent, or that occasional dispersal by common reptiles maintains populations. In contrast with common species, fewer rare species were captured in remnants compared with the Conservation Park, implying that some species may be entirely excluded from the remnants. Our study suggests that the spatial configuration and condition of the fragments sustain populations of many common reptile species. Remnants will therefore be invaluable as attempts are made to restore landscape-scale permeability. However, additional conservation effort should be made to restore plant species that have been lost from the agricultural landscape. Future research should aim to better define the suite of reptile species that may not be able to use the remnants at all.