A decision support tool was created to estimate the treatment efficiency of an Australian drinking water treatment system based on different combinations of extreme weather events and long-term changes. To deal with uncertainties, missing data, and nonlinear behaviours, a Bayesian network (BN) was coupled with a system dynamics (SD) model. The preliminary conceptual structures of these models were developed through stakeholders' consultation. The BN model could rank extreme events, and combinations of them, based on the severity of their impact on health-related water quality. The SD model, in turn, was used to run a long-term estimation of extreme events' impacts by including temporal factors such as increased water demand and customer feedback. The integration of the two models was performed through a combined Monte Carlo–fuzzy logic approach which allowed to take the BN's outputs as inputs for the SD model. The final product is a participatory, multidisciplinary decision support system allowing for robust, sustainable long-term water resources management under uncertain conditions for a specific location.