The scale-up prediction of the model framework of Gahn and Mersmann (1999a,b) has been investigated on the basis of experiments at three different scales. The results show that the seeded batch experiments at 2-I scale were not useful for the estimation of a valid kinetic model of the crystallisation process. The model parameters estimated on the basis of a 75-I DT crystalliser data are capable of predicting certain extent the steady state quantiles of an 1100-I DTB crystalliser. However, fine tuning of the model parameters is still needed to get a good description of the final product quality. (C) 2012 Elsevier Ltd. All rights reserved.