In recent years, more and more water supply systems have become over stressed due to population growth and other economic and social factors. This has led to a need for more efficient operation of policies. One critical aspect in the optimal operation and management of municipal water supply systems is the accurate prediction of the short-term water use. In this paper the relatively new approaches of artificial neural networks and expert systems have been used to forecast daily municipal water use. In the current study four different artificial intelligence models are developed and tested for the city of Lexington, Kentucky. The developed models include a simple artificial neural network model, two rule based expert system models and a combined rule based expert network model. The results of current study indicate that artificial neural networks and expert system technologies have great potential in the area of demand forecasting for complex systems.