Abstract: The fully mechanized mining face is among the most critical areas for dust pollution within a mine, in large mining height faces, the elevated coal fall height during the mining process facilitates the dispersion of particulate matter from the coal cutting machinery into the pedestrian walkway, thereby posing a significant health risk to miners. To elucidate the dust control mechanism of coal buffle in large mining faces during the shearing process, a three-dimensional numerical model was constructed using 15211 working face of Zhangjiamao Coal Mine in northern Shaanxi as a case study. Computational Fluid Dynamics-Discrete Phase Model (CFD-DPM) numerical simulation techniques were utilized to analyze the airflow characteristics, turbulence distribution, dust transport trajectories, distribution range, and particle size distribution under varying coal baffle heights. The findings reveal that an increment in the coal baffle height intensifies its air flow partitioning effect, resulting in a reduction of airflow streamlines in the pedestrian walkway, a progressive decrease in wind velocity, and a precipitous increase in wind velocity near the drum, exceeding 2.5 m/s. When the height of the coal buffle reaches 1.6 meters, a 30-meter-long high wind speed area is formed in front of the coal shearer. The combination of enhanced airflow effect and physical barrier effectively changes the migration path and distribution pattern of dust, significantly suppresses the lateral diffusion of dust, and shortens the exposure time and migration distance of dust. The dust pollution area on the inner side of the sidewalk has been significantly reduced, with the dust mass concentration dropping to 50 mg/m³. Most of the dust is effectively confined to the outer side of the sidewalk, and the interception effect on dust with a particle size greater than 40 μm is particularly obvious. By comparing the dust concentration distribution when the coal buffle was 0.8 m and 1.6 m, the dust concentration at the breathing zone of the sidewalk decreased significantly, with a dust reduction rate of 50.2%. The numerical simulation results were compared with the on-site measured data, and the average error was 6.9%, further verifying the accuracy and reliability of the simulation.