This paper addresses the issue of controlling nonlinear processes by the use of the nonlinear model predictive control formulation. To handle the nonlinearities, a neuro-fuzzy process model is suggested as a means to model processes with strong nonlinearities depending on the operating region. In this paper the neuro-fuzzy approach is used for the modelling and control of a strongly nonlinear pH neutralization process, both in the face of set-point changes and in the face of unmodelled disturbances.