Automation of water treatment plant processes is difficult as a result of complex chemical reactions and physical phenomena. It is further complicated by the dramatic variation which can occur in raw water quality and the lack of general algorithms which describe many of the processes. As a result, the use of traditional process control techniques has had limited success in many water treatment applications. Presented is a preliminary attempt at process automation through a process control system utilizing artificial neural network (ANN) models. This project has shown that the ANN approach is effective at modeling the coagulation process for both turbidity and organics removal using raw water quality and operating conditions. Proposed is an ANN model-based process control that allows easy integration with supervisory control and data acquisition (SCADA) systems. Preliminary results of ANN on-line process control at the Rossdale Water Treatment Plant on the North Saskatchewan River are presented. The ANN control system can be used as a control system for daily operations and can aid operators in selecting chemical doses and operating conditions. The system can also act as a virtual training platform for new operators. Includes 11 references, tables, figures.