Active management is essential to the survival of many threatened species globally. Captive breeding programmes can play an important role in facilitating the supplementation, translocation and reintroduction of wild populations. However, understanding the genetic dynamics within and among wild and captive populations is crucial to the planning and implementation of ex situ management, as adaptive potential is, in part, driven by genetic diversity. Here, we use 14 microsatellite loci and mitochondrial Control Region sequence to examine the population genetics of both wild populations and captive colonies of the endangered warru (the MacDonnell Ranges race of the black-footed rock-wallaby Petrogale lateralis) in central Australia, to understand how historical evolutionary processes have shaped current diversity and ensure effective ex situ management. Whilst microsatellite data reveal significant contemporary differentiation amongst remnant warru populations, evidence of contemporary dispersal and relatively weak isolation by distance, as well as a lack of phylogeographic structure suggests historical connectivity. Genetic diversity within current captive populations is lower than in the wild source populations. Based on our genetic data and ecological observations, we predict outbreeding depression is unlikely and hence make the recommendation that captive populations be managed as one genetic group. This will increase genetic diversity within the captive population and as a result increase the adaptive potential of reintroduced populations. We also identify a new site in the Musgrave Ranges which contains unique alleles but also connectivity with a population 6 km away. This novel genetic diversity could be used as a future source for supplementation.