This research focuses on investigating the effects of the three different grinding aids, consisting of a mixture of amine, glycol and polyol in different ratios, on the model parameters of a two-compartment cement ball mill and an air classifier. Within the content of this work, sampling campaigns were organized around a cement grinding circuit and each of the three grinding aids at varying dosage rates were tested. Then the ball mill and the air classifier were modelled by using perfect mixing approach and Whiten's equation respectively. The relations between the types and dosage rates of grinding aids and model parameters were examined. The fact that such an examination has not been made previously by using industrial data, which makes this research unique. It was deduced that depending on the chemical composition of grinding aids, their effects on model parameters were varied and ultimately all of them improved the performance of the grinding and the classification operations. (C) 2018 Elsevier B.V. All rights reserved.