A method of optimal experimental design for parameter estimation in unstructured growth models is presented. The approach is based on a method suggested by Munack (1991) for application in fed-batch processes. In a critical analysis of this method, special emphasis is given to the model validity, because unstructured growth models often are not valid under transient conditions. In consequence, a combined object function has been introduced, which considers model validity and the accuracy of the kinetic parameters to be estimated. The application of this method for fed-batch processes leads to satisfactory results. Investigations of different fed-batch strategies regarding model validity and the quality of parameter estimation are presented. In addition, an experimental verification has been performed with fermentations of the yeast Trichosporon cutaneum.