Despite many attempts to establish material grindability in a milling process, it remains very difficult to predict the milling performance. In this study, impact milling tests were carried out under varying operational conditions in a UPZ100 impact pin mill. The product size distribution and fineness of milled alumina particle are reported and analyzed. A multiscale framework coupling discrete element method (DEM) and population balance model (PBM) is proposed to predict the milling performance of the mill. The impact velocity and impact frequency information from DEM is utilized to inform the mill operation parameters of the PBM model at the process scale, whilst the material grindability parameters are evaluated using constrained optimization to match a milling lest. The predictions of the product size distribution show very good agreement with the experimental results. This indicates that the multiscale model is promising for optimizing the design and operation of mills. (C) 2020 Published by Elsevier B.V.