Abstract: Turfy soil is a special kind of humus soil formed by oxidation and partial decomposition of plant residues in the environment of waterlogging. Turfy soil contains a lot of plant fibers, and is with adverse engineering geological characteristics such as high permeability, high pore ratio and high-water content. In order to investigate the pore characteristics of the turfy soil and the percolation law, CT sequence images of turfy soil in Dunhua, Jilin Province are obtained. The optimal threshold of image binarization is determined using the mask volume calculation of Mimics and porosity of turfy soil. The 3D reconstruction model is constructed. Meanwhile, based on the Lattice Boltzmann Method principle, the PALABOS code is adopted to conduct the single-phase seepage simulation of the turfy soil. In addition, the differences of simulated permeability values and the seepage field of the soil under constant pressure are studied. The distribution of flow lines and flow velocity within the pore space are analyzed. The results show that internal pores in turfy soil vary in size. The large pore size and small pore size exhibit different patterns. The distribution of plant fibers and root-like pores are the main reason for the formation of large pore size channels in turfy soil. Therefore, the permeability of turfy soils is mainly affected by the content of plant fibers in the soil and the degree of decomposition. The more content of plant fiber, the lower the decomposition degree is and the higher the permeability of turfy soil is, and vice versa.