This paper proposes an adaptive control scheme for a fermentation process. This incorporates identification of the process parameters via a recursive identification in which a non-linear process model is parameterised in terms of a linear identification model. This allows the application of traditional least-squares techniques for on-line identification of system parameters. As the primary states, biomass and substrate, are not directly measurable on-line state estimation via the method of Kalman filtering is proposed. The filter algorithm uses the measurements from an established sensor for gaseous carbon dioxide to estimate process states. Adaptive control of the process using a pole placement state feedback controller is utilized to give the process some desired dynamic response characteristics. Results are presented demonstrating parameter convergence, estimated state trajectories and regulation of the system output and conclusions are drawn which summarise the successful approaches to fermentation control and make some recommendations on the choice of algorithms and direction of future work.