Membrane fouling is one of the important concerns of membrane technology for water treatment. The purpose of the study was to develop a simple flux model to estimate the characteristic fouling times for ultrafiltration systems using based on operational data. A simple first order model was developed to correlate flux to characteristic system parameters. The model was verified with data from a pilot plant. The model parameters were estimated using a genetic algorithm developed at the Florida International University. The genetic algorithm predicted the calibration parameters and characteristic fouling times for individual runs between 23 and 129 generations.