The gas-solid two-phase flows in fibrous filters were simulated by computational fluid dynamics (CFD) technology. The pressure drops and filter efficiencies with different operating conditions and geometry parameter, including face velocity, particle size, and solid volume fraction (SVF) were calculated. The effects of the operating conditions and geometry parameter on the filter performance of the fibrous filter were obtained. The results indicate that the pressure drop increases linearly with the face velocity and the predicted values of the pressure drops are in excellent agreement with the experimental correlation. Filtration efficiency decreases with the face velocity for submicrometer particles (0.1 mu m) and, for larger particles (1 mu m) the tendency is just the opposite. The filtration mechanism is different for different particle sizes. For the filter in this paper, when the particle size is smaller than 0.2 mu m, Brownian diffusion plays a significant role in the filtration process. When the particle size is greater than 0.5 mu m, inertial impaction becomes an important capture mechanism. For particle sizes In the range of 0.2-0.5 mu m, the Brownian diffusion and inertial impaction are both relatively weak and, therefore, the filtration efficiency has the least value in this range. Additionally, the SVF distribution is an important geometry parameter in the filter. The filtration efficiency of the filter with a decreased SVF (geometry B) along the thickness of the filter is higher than that of the filter with the even SVF (geometry A), while maintaining a low pressure drop.