Model-based control of bioprocess is a difficult task due to the challenges associated with bioprocess modeling and the lack of on-line measurements. In this study, a robust controller using minimal process knowledge and minimal measurement information is developed and applied to the ethanol regulation in cultures of Saccharomyces cerevisiae. A Youla parametrization is selected in order to reject the disturbance associated to the substrate demand for cell growth and to robustify the control scheme to unstructured uncertainties and measurement noise. The performance of this control scheme is illustrated in simulation and with real on-line experimental data. (C) 2006 Elsevier Ltd. All rights reserved.