Abstract: In many industrial fields such as mineral mining, metal smelting, and wood processing, the efficient capture and treatment of dust pollutants has always been a common technical challenge across industries. Traditional dust capture and removal technologies frequently face performance limitations including uneven flow field distribution, low particle capture efficiency, and high energy consumption. Motivated by these challenges, this review systematically examines recent advances in the application of experimental methods and Computational Fluid Dynamics (CFD) in the flow field simulation and structural optimization of dust capture and removal systems. It identifies the key parameters that are critical to overall system performance and analyzes the application of machine learning in optimization. CFD technology has become a core tool for gaining deep insight into the internal mechanisms of dust capture and removal processes, as well as for driving structural innovation and performance enhancement. There still exist some problems including insufficient simulation accuracy, lagging dynamic control, the limitations of single-technology approaches, and poor model generalizability in the current dust capture and removal system. To address these problems, future research should focus on developing non-spherical particle multiphase flow models, digital twins and real-time closed-loop control, low resistance and high-efficiency composite technology, and modeling approaches that integrate physical mechanisms with data-driven methods.