In this work, an optimum experimental design strategy has been developed to increase the reliability of chemical reaction engineering models when few experiments are to be performed. The strategy has been used for heterogeneous liquid-liquid reactions carried out in a batch reactor. The specific case of alkaline hydrolysis of n-amylacetate has been selected to illustrate the proposed methodology, allowing experiments to be carried out to identify, the associated kinetic parameters. The kinetic parameters that arise from the model system of ordinary differential equations have been estimated by weighted least squares and a D-optimum sequential design has been implemented to reduce the volume of the joint confidence region of the parameters. Estimates and their uncertainty obtained at the end of the strategy, as well as the generalization ability of the prediction model, are presented and discussed. (c) 2005 American Institute of Chemical Engineers.