Dispersal is a critical aspect of the biology of most organisms, but movement patterns are poorly known for most animals. Knowledge of movement patterns allows biologists to construct realistic spatial models, either for theoretical or empirical purposes. Cane toads (Bufo marinus) have high survival rates (30–70% per annum in Australia), but very low recapture rates over long periods in marking studies, suggesting that they may be nomadic. Here we model the movement patterns of cane toads to determine whether they move nomadically. Using radiotracking, we quantified the movements of 64 cane toads, at two widely separated locations in northern Australia. We established that toad movements were indeed nomadic by comparing real movements with those generated by an individual-based, correlated random walk model that selected randomly from empirical distributions of movement parameters. In the wet season, at both locations, real movements were similar to those generated by the model, indicating that toads moved nomadically during this time of year. During the dry season at both locations, toads used a smaller area and travelled less straight paths than predicted by the model. Nomadic movements were more common and extensive during the wet season, although some individuals moved nomadically at all times of year. To determine whether the model provided a good description of toad movements over the long term, we compared observed disappearance rates from a mark-recapture study conducted in one of our study sites, with disappearance rates predicted by the model. There was good agreement between these two methods of measuring toad movements. Previous studies indicated that food availability was lowest and desiccation risk and predation risk were highest in the dry season, suggesting that nomadic movement of toads is most extensive when the costs of movement due to desiccation, predation and energy limitation are lowest.