Dense medium cyclone (DMC) is widely used to upgrade run-of-mine coal in modern. coal preparation plants. The flow within it is very complicated, with multi -phases involved including air, water, coal and magnetic/non-magnetic particles of different sizes, densities and other properties. In this work, the effect of coal particle density distribution, one of the most important variables and highly related to coal type, is systematically studied using a combined approach of Computational Fluid Dynamics (CFD) and Discrete Element Method (DEM). In particular, the so called Johnson's S-B function, which can describe a wide range of distributions, is employed to represent different particle density distributions. For a given density range, the function is characterized by two parameters: median particle density p(0.5) and distribution parameter 03, with the latter describes the spread of the distribution. The effects of the two parameters on the flow and performance of a typical DMC are quantified. Moreover, their implication to practical operation, e.g. when the amount of near gravity material is high, is discussed. The results are also analysed in terms of the medium and particle flow fields, particle-fluid, particle particle and particle-wall interaction forces to understand the fundamentals of the operation. The results obtained in this work should be useful to the design and control of DMC operations with different coal density distributions or coal types. (C) 2016 Elsevier B.V. All rights reserved.