To remove the fouling layer and thus to maintain an economically feasible net permeability, air scour of the membrane by coarse bubbling and backflushing from the permeate side (hollow-fibre modules) or relaxation (flat sheet modules) are currently employed. These measures are neither optimised nor controlled, i.e., they are often carried out before they are necessary, thereby wasting energy, permeate or chemicals, or too late. This leads to losses in productivity and to environmental hazards through the fori-nation of chemical cleaning by-products such as AOX. Mechanistic models which describe filtration and fouling mechanisms exist, but have not yet been used for process control. The aim of this work is to develop model-based control strategies for membrane operation which reduce energetic expenditure and increase filtration efficiency. A novel model-based control strategy for membrane operation is presented which is based on an automated recognition of the currently dominant filtration mechanism. Thus, suitable control actions can be carried out in order to increase the permeability respective to each mechanism improving the filtration performance in membrane bioreactors (MBR). The validation of the model recognition approach is demonstrated using experimental data from a test cell as well as from an MBR pilot plant.