Mixing of cohesive micro and nano-powders is difficult because they form large agglomerates due to the high interparticle forces. In order to better understand the mixing of cohesive particles, discrete element method (DEM) based modeling was performed for the magnetic assisted impaction mixing (MAIM), which is a high shear mixer previously shown to be capable,of mixing at the nanoparticle scale. The JKR cohesion force model was used to represent interparticle cohesion. Agglomerates were formed based on the surface energy of individual particles, thus better capturing the effect of cohesion on the initial state. The effects of magnet-to-sample mass ratio, magnet size and surface energy of non-magnet particles on the homogeneity of mixing (HoM) were investigated. Simulation results show that the mixing will be faster with smaller magnet sizes at fixed mass ratio, by increasing the mass ratio, or by decreasing the surface energy; the latter had a significant effect on the process of mixing. When non-magnetic particles had higher surface energy, homogeneous mixing required longer processing times since higher collision numbers and collision energies were necessary to deagglomerate the particles. Results show that when the collision energy between magnets and non-magnets exceeds the cohesive energy, the mixing would reach a steady state at shorter processing intervals. The results qualitatively agree with previously published results, suggesting that this system model, which involves the formation and utilization of agglomerates in simulations, is applicable to cohesive powder mixing. (C) 2013 Elsevier B.V. All rights reserved.