Anecdotal evidence suggests toxin producing cyanobacteria are becoming more abundant in drinking water reservoirs around the world. In Queensland, the three most common toxin producing cyanobacteria found in freshwater storages are Microcystis aeruginosa, Anabaena circinalis and Cylindrospermopsis raciborskii. Blooms of these species have public health implications due to their known potential for the release of hepatotoxins (M. aeruginosa and C. raciborskii), and neurotoxins (A. circinalis). This paper will present initial results from a study which aims to predict the effects of climatic changes on the occurrence and distribution of these potentially harmful cyanobacteria. The results from examination of historical limnological data will be used and integrated with predictive climate models to assess potential future risk from cyanobacterial blooms. The application of an Artificial Neural Network as a modelling and predictive tool of cyanobacterial blooms in reservoirs will be examined. Such systems have proved useful in the past for prediction of bloom events, which allows time for the implementation of strategies to reduce the impact of cyanobacterial toxins on public water supplies. The information from the modelling approaches has provided the framework for experimental investigations incorporating mesocosm and laboratory culture experiments.