A bead mill was used to disperse AlN nanoparticles into coolant-oil to make thermally efficient nanofluids. This apparatus disperses particles by imparting direct impact to the agglomerates of nanoparticles using beads of certain size circulated by added centrifugal forces in the milling vessel. A mathematical model based on population balance was developed to determine optimum operating conditions of the bead mill process. A dimensional analysis for the process model led to two dimensionless groups eta(c) and eta(b), whose ratio (eta(c)/eta(b)) represents the relative intensity of the coagulation and breakage phenomena taking place in the mill. Numerical simulations were performed to investigate the effect of process operation variables upon the particle size distribution, and three adjustable model parameters were selected and estimated through regression analysis based on a portion of the experimental data. The population balance model with the estimated parameters yielded prediction results that reasonably matched the rest of the experimental data obtained under different operating conditions, hence is expected to be used as a tool for finding optimal milling conditions. (C) 2012 Elsevier B.V. All rights reserved.