The high costs associated with many fermentation processes in an increasingly competitive industry make the optimisation of bioreactor performance very desirable. Attempts to improve performance have primarily been focused on the development of online adaptive methods. Although these techniques can provide the optimum solution, each had limitations in practical implementation. The paper addresses these problems and considers their effects on the optimisation of bioreactor performance. An online time series model is sought in order to evaluate the steady-state performance of the bioreactor. A simple algorithm is developed around this theme and is analysed to show convergence to the optimum solution. Nonlinear simulation results, using the Monod equations for yeast fermentations, are presented. For this processes, two approaches using an online identified linear model or nonlinear model, are shown to give optimum results.