Abstract: The filter layers of tailings dams and earth-rock dams are key to dam seepage stability. Long-term seepage drives fine particle migration, causing piping or clogging and altering permeability. Existing studies analyze piping or clogging separately, lacking quantitative evaluation of seepage history’s impact on soil permeability. This study investigates gap-graded soils through an integrated approach combining laboratory seepage tests, CFD-DEM fluid–solid coupling numerical simulations, and CT image analysis. A predictive formula for Hydraulic Conductivity considering fine particle content is derived, and modified permeability formulas applicable to clogged soil and piping soil are proposed. The results indicate that: 1) under identical gradation, dry density, and void ratio, the Hydraulic Conductivity exhibit a distinct pattern: piping soil > natural soil > clogging soil; 2) during piping, fine particles become suspended and form preferential flow channels, increasing permeability by 2.73 times compared to natural soil; during clogging, fine particles deposit to form a low-permeability clogging layer, reducing permeability by more than 50% relative to natural soil; natural soil, with its uniform particle distribution, shows intermediate permeability; 3) numerical simulations and CT analyses reveal that the non-uniform reconstruction of pore structure due to fine particle migration—manifested as an increased proportion of large pores in piping soil and concentrated small pores in the upper layer of clogging soil—constitutes the intrinsic mechanism for permeability differentiation. The modified formulas show a 40%~60% improvement in agreement with measured values compared to traditional homogeneous models. This study deepens the understanding of the mechanism by which seepage history influences soil permeability and provides theoretical and technical support for precise prevention and control of seepage failure in hydraulic and geotechnical engineering.