This document describes a study applying a recursive orthogonal least squares (ROLS) algorithm to develop a neural network model for a continuous fermenter system using a nonlinear model predictive control (NMPC) approach. The ROLS algorithm is used to update the weighting matrix to model the nonlinear, multivariable system dynamics. The NMPC controller using this model is tested on the fermenter by changing setpoints. It is shown to reach new setpoints faster than alternative controllers like IMC, PI, and FLC controllers.