Ethanol concentration during fed-batch growth of yeast on sucrose as substrate is to be controlled. Rapid implementation and flexible adaptation to changing process characteristics are of major importance. Therefore the process model is established in view of control, including only essential process features relevant to this aim. This leads to a low number of model parameters which are estimated on-line. No off-line measurements are required. Based on this model, a predictive control algorithm provides the future manipulated variable by analytical calculation, avoiding numerical optimisation methods. Hereby calculation time is significantly reduced. Thus all the software for data-acquisition, supervision and control can be run on one PC for several bioreactors at the same time. The controller software was implemented together with additional supervisory routines on an industrial DCS. Several experiments were carried out with different strains of Saccharomyces cerevisiae. The controller adapts very well to the characteristics of the respective strain and tracks the setpoint profile precisely.