Kinetics is essential for chemical reactor modelling, in particular to reduce the inherent risks of extrapolation going along with scaling-up. Pharmaceutical industries are especially concerned. However, when chemical systems are very complex, development of good models may lead to prohibitively expensive and time consuming experiments. The aim of this paper is to describe an efficient experimental design strategy for discrimination of stoichiokinetic models. The proposed methodology is based on model-based experimental design (optimal design), which uses information already acquired on models to determine the best conditions to implement a new experiment with the highest discrimination potential. The combination with microreactor technology is also proposed in this work. The whole procedure for model discrimination is firstly described in detail and then, applied to a numerical study case, consisting of a chemical synthesis carried out in a microreactor. The discrimination procedure efficiently leads to the determination of the single adequate model among the various potential models proposed before the implementation of the designed experiments. It is verified that the procedure does not depend on the set of preliminary experiments and is time-saving when compared to a classical factorial plan. (C) 2016 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.