Abstract: To address the suboptimal efficiency of respirable coal dust capture in underground coal mines, this study proposes a multi-stage stepwise supersonic dynamic micro-fog collaborative dust removal fan by integrating supersonic coaxial pneumatic atomization technology. The dust removal unit structure was designed based on step function principles. Using the DPM method combined with k-ε turbulence model, we investigated the internal flow fields, droplet dynamics, and particle trajectories within different step structures. Experimental tests were conducted using thermal anemometers, laser particle analyzers, dust samplers, and dispersion testers to characterize airflow velocity, droplet size distribution, and their evolution patterns across step structures. The optimal structure and dust removal mechanism were identified. Results demonstrate that multi-stage step structures exhibit abrupt changes in lateral airflow resistance due to airflow separation, vortex generation, and wall impingement, forming dynamic competition with energy accumulation effects. Airflow velocities in Structure 5 fluctuate slightly around 2 m/s, while other structures show ascending trends from levels 1 to 5. Droplet sizes across different structures undergo rapid coagulation under extensive vortex and turbulent centrifugal effects, increasing from 13.24−16.47 μm to 17.12−20.36 μm along migration paths. The heterogeneous turbulent field within multi-stage step structures effectively prolongs the dust-fog coupling duration, with coupling efficiency significantly influenced by step structure design. Compared to Structure 5, other step structures achieved maximum efficiency improvements of 211.93%,42.7%, and 20.93% for particle sizes 0−2.5 μm, 2.5−5 μm, and 5−10 μm, respectively. The symmetrical reverse-order double-step structure achieves peak performance in both vorticity and turbulent kinetic energy. The high-speed airflow rotation and intense turbulence synergistically enhance the dust-fog coupling effect. Under this stepped structure, the coupling characteristics between fog droplets and dust particles exhibit exceptional compatibility, resulting in optimal dust removal efficiency of 96.97%. The efficiency breakdown across particle size ranges is: 93.04% for 0−2.5 μm, 92.14% for 2.5−5 μm, 85.13% for 5−10 μm, 87.83% for 10−20 μm, and 98.48% for 20−50 μm. These findings provide theoretical and technical support for efficient control of respirable coal dust in underground coal mines.