Understanding animal movement patterns is fundamental to ecology, as it allows inference about species? habitat preferences and their niches. Such knowledge also underpins our ability to predict how animals may respond to environmental change, including habitat loss and modification. Data-logging devices such as GPS trackers and accelerometers are rapidly becoming cheaper and smaller, allowing movement at fine scales to be recorded on a broad range of animal species. We examined movement patterns of an arboreal mammal (bobuck, Trichosurus cunninghami) in a highly fragmented forest ecosystem. The GPS data showed males travelled greater distances than females in linear roadside strip habitats, but not in forest fragments. The accelerometer data showed that both sexes exhibited higher activity levels in roadside habitats compared to forest fragments. By coupling GPS and accelerometer data, we uncovered for this species an ecological pattern similar to other mammals: that male bobucks had higher activity levels than females for a given distance travelled. Our findings also suggest that habitat fragmentation changes the amount and type of activity bobucks perform while moving, and that linear forest strips could be considered "energetically challenging" habitats, which informs how we should manage the spatial distribution of key supplementary resources for this species such as nest sites and minimum fragment sizes.