A novel fluorescence-based approach is proposed for modeling, predicting, and optimizing different fouling dynamics in an ultrafiltration (UF) process for drinking water treatment. Principal component analysis (PCA) was used to extract information in terms of principal components (PCs), related to major membrane foulant groups, from fluorescence excitationemission matrix measurements captured during UF of natural river water. The evolution of PC scores during the filtration process was then related to membrane fouling using dynamic balances of latent variable values (PC scores). This approach was found suitable for forecasting fouling behaviors with good accuracy based solely on fluorescence data collected 15 min from the start of the filtration. The proposed approach was tested experimentally through model-based optimization of backwashing times with the objective of minimizing the energy consumption per unit amount of water produced during the filtration process. This approach was also useful for identifying fouling groups contributing to reversible and irreversible fouling. (c) 2011 American Institute of Chemical Engineers AIChE J, 2012