In this study, the perfect mixing model was used to investigate its accuracy level, under different laboratory conditions, in predicting the particle size distribution of industrial ball mills discharges. For this purpose, data sets of two laboratory ball mills with eight different compositions of balls and two industrial mills of a copper processing plant for seven different tonnages, which totally included 56 simulation operations, were used. For simulation, the necessary data were obtained through performing the breakage distribution function and kinetic grinding tests using laboratory mills. The results were used to determine the first order grinding kinetics and normalized breakage rate parameters. For the industrial scale, the simulation process was carried out using data, perfect mixing model equations and JKSimMet software. The results showed that the operating conditions of the laboratory mills were quite affected by the predictive power of the desired model. Comparing the measured and simulated values of P80, it is clear that 2 minutes of first order grinding using the Bond laboratory ball mill with standard operating conditions and single size ball load of 20 mm provided the best prediction with trivial errors, less than 10%, for all seven tonnages of the industrial mills. The results of this study together with more investigations on different plants can be helpful in optimization, simulation and scale-up procedures of ball mills.