A process diagnosis method for membrane water treatment plant was developed using a constant flux membrane fouling model. This diagnosis method can be applied to a real-field membrane-based water treatment plant as an early alarming system for membrane fouling. The constant flux membrane fouling model was based on the simplest equation form to describe change in trans-membrane pressure (TMP) during the filtration cycle from a literature. The model was verified using a pilot-scale microfiltraton (MF) plant with two commercial MF membrane modules (72 m2 of membrane area). The predicted TMP data were produced using the model, where the modeling parameters were obtained by the least square method using the early plant data and modeling equations. The diagnosis was carried out by comparing the predicted TMP data (as baseline) and real plant data. As a result of the case study, the diagnsis method worked pretty well to predict the early points where fouling started to occur.