Flies are important arthropod pests in intensive animal facilities such as cattle feedlots, with the potential to cause production loss, transmit disease and cause nuisance to surrounding communities. In the present study, seasonal population dynamics of three important nuisance flies, namely house flies (Musca domestica L.), bush flies (M. vetustissima Walker) and stable flies (Stomoxys calcitrans L.) (Diptera: Muscidae), were monitored on cattle feedlots in south-eastern Queensland, Australia, over 7 years. Musca domestica was by far the dominant species, comprising 67% of the total flies trapped. Models were developed to assess the relationship between weather parameters and fly abundance and to determine whether population trends could be predicted to improve the timing of control measures. For all three species, there were two main effects, namely time-of-year (mainly reflected by minimum temperatures and solar radiation) and rainfall. The abundance of all three species increased with increasing temperature and rainfall, reaching a peak in summer, before decreasing again. Rainfall events resulted in significantly elevated numbers of M. domestica for up to 5 weeks, and for 1 week for M. vetustissima. Peak fly numbers were predicted by the model to occur in spring and summer, following 85–90-mm weekly rainfall. The population dynamics of S. calcitrans were least influenced by rainfall and it was concluded that weather variables were of limited use for forecasting stable fly numbers in this environment and production system. The models provide a useful tool for optimising the timing of fly-control measures, such as insecticide or biopesticide applications, adding to the efficiency of integrated control programs.