The predictive control model is an advanced control technique which determines a future control profile based on predictions made according to a process model to match a given performance criterion. This technique was developed at the end of the 1970s and implemented with huge success in the petrochemical and refinery industry. At the end of the 1980s process regulation engineers working in the water treatment industry became interested in alternative methods to the feedback control used at that time. This paper describes the development of a feed forward based control algorithm for the regulation of a fullscale ozonation process. The manipulated variable is the ozone production (ozone generator power regulation) and the target (set point) is the disinfection goal expressed as a specified CT value. On-line data (water quality and plant operation characteristics) is exported to a PC spreadsheet. Field and lab data (water ozone demand, ozone decay constant) is entered manually. Database tables are developed (CT values as function of water temperature and for specific disinfection goals). Summary reports and trend-charts are prepared, evaluated and modified to address specific situations. Includes 7 references, tables, figures.