Lysine, an important nutrient in dairy formulations, is easily damaged during spray drying. The quality of spray dried dairy powders can be enhanced when lysine loss kinetics are coupled with drying kinetics and computational fluid dynamics of spray. Thin-film drying in a convective drying channel was applied to study the drying kinetics of the model dairy-formulation. The drying kinetics was modeled with the reaction engineering approach which was successfully implemented in the computational fluid dynamics (CFD) simulation of spray drying. Lysine loss was modeled by reaction kinetics taking into account the temperature, water content and physical state of lactose. A 3D CFD model of spray drying was set up to determine the particle properties along their residence time. The model was validated with the experimental results of spray drying on a pilot scale. A good agreement between the experiments and the simulation was obtained. The reaction kinetics model was coupled with the particle properties along the particle tracks to create a predictive tool for the lysine loss. Thus, as it was hypothesized from the spray-drying experiments, the importance of the particle residence-time was highlighted. The following optimization study revealed potential strategies to ameliorate the spray drying process. In this context, the dryer geometry, air flow pattern and droplet atomization should be reconsidered carefully. The proposed approach can be transferred to other spray drying applications, e.g. for the targeted formation of microcapsules or to enhance the survival of microorganisms. (C) 2015 Elsevier Ltd. All rights reserved.