Considering being applied to industrial applications with billions of particles, the discrete element method (DEM) is still limited by the computational competence of computers, regardless of the advances in computer hardware or parallel programming techniques. In this paper, therefore, a coarse-grain (CG) discrete element method is adopted to deal with the immense number of particles in granular mixing, which has the potential to gain considerable computational advantage for industry scale DEM simulations. The experiment results as well as the simulation results prove the validity of taking place of smaller particles by larger particles with sacrificing a little detail in a drum flow. The simulation results also indicate that the rotation speed of the drum nearly has no effect on the application of the CG-DEM. The mixing tendency as well as the flow pattern of particles in the CG system is quite similar to the original system. In addition, the simulation results of an industrial application show that the CG-DEM is applicable to the double-screw conical mixer. The granular mixing in the double-screw conical mixer can be predicted by the CG-DEM, and the prediction of device power consumption is accurate by correcting the friction coefficient based on the theory of the CG-DEM, and the prediction of the wear rate of the device is broadly accurate after being corrected by a correction factor. Using CG-DEM can efficiently save computational cost with cutting back on the level of detail and accuracy. (C) 2019 Elsevier B.V. All rights reserved.