Knowledge of internal variables in horizontal rotating drums is useful for the optimisation and control of industrial processes including milling, mixing, coating and agglomeration. The energy of particle-particle (p-p) impacts, which is the focus of this paper, is a variable of particular interest because of its relevance to the efficiency of these processes. For example the comminution rate in a grinding mill or the structural and chemical transformations when alloying in mills are both determined by the impact energy of particles. There are two main obstructions that prevent the measurement of internal variables in horizontal rotating drums: the harsh internal environment, which makes internal installation of sensors for online measurement impractical, and the lack of sensors that can directly measure the internal variable of interest. The above difficulties warrant the implementation and development of soft-sensors to predict the particle-particle impact energy from variables that can be physically measured outside the drums. Based on a discrete element method (DEM) model, such a soft-sensor model is developed for horizontal rotating drums that can quantitatively predict p-p impact energy within 5% of its actual value. Crown Copyright (c) 2009 Published by Elsevier B.V. All rights reserved.