Abstract: The multi-model predictive control strategy is presented to solve the control problem of typical nonlinear systems, i.e., the wastewater treatment process. The way to develop the multi-model is based on the local models which were systematically developed via finding out the data set from numerous accumulated system input/output database to match the current modality and taking advantage of the fitting of polynomial. Then, the multiple model were developed according to the variation in system modality. Combining the multiple models thus developed with predictive control, the multi-model predictive control strategy is proposed, thus solving the control problem of a class of unknown-structure nonlinear systems just on the basis of numerous historical database. The simulation test illustrates the validity of this approach.