Identifying explicit hypotheses regarding the factors determining genetic structuring within species can be difficult, especially in species distributed in historically dynamic regions. To contend with these challenges, we use a framework that combines species distribution models, environmental data and multi-locus genetic data to generate and explore phylogeographic hypotheses for reptile species occupying the coastal sand-dune and sand-plain habitats of the south-western Australian biodiversity hotspot, a community which has both a high diversity of endemics and has varied dramatically in spatial extent over time. We use hierarchical amova, summary statistic and distance-based analyses to explicitly test specific phylogeographic hypotheses. Namely, we test if biogeographic vicariance across barriers, habitat stability, population isolation along a linear habitat or fragmentation across different environments can explain genetic divergence within five co-distributed squamate reptile species. Our results show that patterns of genetic variation reflect complex and species-specific interactions related to the spatial distribution of habitats present currently and during repeated glacial minima, as opposed to being associated with historical factors such as habitat stability between glacial and inter-glacial periods or vicariant barriers. We suggest that the large impact of habitat characteristics over time (i.e. relative levels of habitat connectivity, climatic gradients and spatial heterogeneity of soil types) reflects the ecological restrictions of the sand-dune and sand-plain reptile communities and may explain the lack of concordance across taxa. The study demonstrates the general utility of the approach for assemblage-level, as well as single species, phylogeographic study, including its usefulness for exploring biologically informed hypotheses about what factors have influenced patterns of genetic variation.