This study is about a method to control membrane fouling. Developed orifice-based membrane filtration system scraped off the cake layer formed on membrane surface by inducing high shear stress owing to the turbulence generation. Colloidal foulants and organic foulants, among 4 types of foulants, were studied as model cake-forming foulant in membrane filtration. Mathematical force balance was calculated to understand the interaction between particle and membrane. As a final step, artificial neural network was utilized to predict membrane performance of complex input parameters in filtration.