A semi-empirical model that predicts permeate fluxes in high-shear microfiltration systems is proposed. The model assumes that non-diffusive transport phenomena are the main mechanisms for the back-transport of particles from the membrane surface to the bulk solution during filtration. It also considers the use of reverse filtration (back-flushing) to control and minimize concentration polarization and fouling. This model incorporates an equation for the transient flux based on a particle mass balance at the membrane surface. The model has been validated by laboratory experiments using two different suspensions: flexographic ink (Phi(b) = 0.005) and yeast (Phi(b) = 0.01), where Phi(b) is the solid volume fraction in the suspension. The membrane system adopted in this research consists of a bundle of hollow-fiber microfiltration membranes submerged in a tank into which the effluent is introduced. Vacuum is used inside the fiber lumen to create transmembrane pressure differential, and aeration to promote high-shear stress at the membrane surface to minimize concentration polarization and fouling. Periodic reverse filtration (back-flushing) is also used. The experimental validation of the model was carried out by substituting the experimentally determined suspension and operating variables into the model. The results showed good agreement between model prediction and experimental observations.