Improving our understanding of resources required by threatened small mammals is directly relevant to the success of habitat restoration and species reintroduction programs. In a case study based on southern brown bandicoots (Isoodon obesulus obesulus; Mammalia: Peramelidae) occupying a remnant of open forest with a sclerophyllous shrub understory, we investigated microhabitat composition using multivariate analysis, and disproportional use of these habitats using a variety of techniques, including principal canonical correlation vectors, chi-square test, compositional analysis (CA), and nonparametric multiplicative regression. Spool-and-line tracking of bandicoots enabled floristic and structural parameters to be recorded from sites of activity and compared with sites randomly located within 5-ha grids centered over each trapping transect. Each of the 4 methods applied contributed useful interrogation of the potential resources required by I. o. obesulus, with most disproportional use of microhabitats across activities detected using CA. Analyses supported fine-scale preference for Xanthorrhoea semiplana–dominated microhabitat across all activities, nesting in Banksia ornata–dominated microhabitat, moving and foraging in Allocasuarina muelleriana subsp. muelleriana microhabitat, impartial use of and increased reliance on burrows for shelter in Eucalyptus cosmophylla open forest with Melaleuca decussata understory microhabitats, and avoidance of Cyperaceae-dominated microhabitat and mixed heath. These results show that within broadly suitable vegetation communities, I. o. obesulus differentially utilizes a mosaic of microhabitats for a range of activities associated with shelter and foraging. Hence, the success of reintroduction and habitat restoration programs may be improved by considering the availability of preferred microhabitats (or alternative structures in suboptimal habitats, e.g., burrows), and protecting or revegetating indicator and associated species of known preferred microhabitats, respectively.