In order to investigate the diffusing pollution patterns of respirable dust coupled with multiple sources in the working regions of Longwall Face, computational fluid dynamics (CFD) approach was adopted based on k-epsilon two-equation turbulent model and the discrete phase model (DPM). The modelled 52,305-1 Longwall Face (Daliuta coal mine, Shaanxi, China) is characterized by three areas division, and respirable dust diffusing pollution was analyzed on a macroscopic and on a mesoscopic scale. The results show that at macro-scale, the localized dust concentration is proportional to the density of dust particles' trace lines. Respirable dust generated from the front drum largely contributed to the pollution of the roof space. The pollution caused by respirable dust generated from the advancing support to the roof space was the most serious, with dust concentrations exceeding 1620 mg/m(3) that were difficult to reduce. Air flow streamlines and trace lines of respirable dust particles coincide at a high degree. At the meso-scale, respirable dust particles near the variable cross section reach the greatest particle velocity (>2.6 m/s.) The particle groups generated from the front and back drum mostly spread towards the footway space for most of the diffusion time, while the particles group originating from the advancing support moves towards the floor during the first 37.5 s. Particles of this latter group reach a maximum average velocity, create a pollution density of the widest range, and also cause the most direct pollution of the footway space. These mesoscopic behaviors verify the movement and concentration distribution characteristics of the respirable dust that were obtained macroscopically. The results provide novel insights on diffusing dynamics of respirable dust particles that also can be applied to other industrial and environmental areas as well. (C) 2018 Elsevier B.V. All rights reserved.