Abstract Membrane processes in water treatment often experience permeate flux decline due to fouling caused by natural organic matter (NOM). A membrane transport model incorporating surface and internal pore fouling was developed for prediction and simulation of permeate flux and concentrations of permeate and concentrate. This model considered resistances due to membrane material, concentration polarization, gel formation, and internal pore fouling. Contact angle measurements employed for evaluating hydrophobicity and membrane fouling potential showed that the polyethersulfone HFK-131 ultrafiltration membranes were more hydrophobic, and therefore, more susceptible to organic fouling than the cross-linked aromatic polyamide NF-70 nanofiltration membranes. Membrane filtration tests were conducted with groundwater containing NOM under various operation conditions to evaluate permeate flux decline patterns due to organic fouling, as well as solute concentration profiles in the permeate and concentrate streams. The predictive capability of the proposed model was excellent as evidenced from comparison of experimental results and model simulations. Model simulation and sensitivity studies provided a well-founded theory based on the dependence of transport resistance components and membrane performance on various input parameters. These parameters included the mass-transfer coefficient, solution dynamic viscosity, membrane resistance, and resistance per unit gel-layer thickness.