Abstract: This study aimed to determine the relationship between different under-forest economy modes and soil moisture permeability in the karst mountain of Southwest China. Soil infiltration and the influence factors of six different under-forest economy modes in Chongqing City were studied, with the pure forest as the control. The study showed significant differences in soil infiltration among different under-forest economy modes and soil layers. Infiltration was highest (449 mL) in 0 10 cm soil layer under the mode of fungi cultivation in Dendrocalamus latiflorus forest. The second highest (427.5 mL) was under the mode of grass cultivation in Ficus lacor forest, followed the soil under the mode of poultry feeding in D. latiflorus forest (389.4 mL). Soil infiltration was poorest (241.5 mL) under the mode of fungi cultivation in Eucalyptus robusta forest. Soil infiltration in each under-forest economy mode decreased with increasing soil depth. The common infiltration model was the most suitable for simulating soil infiltrating processes in the study area, followed by the Kostiakov equation, and then the Horton equation. Soil infiltration evaluation via correlation analysis showed that the strongest soil infiltration capacity was of the mode of fungi cultivation in D. latiflorus forest. This was fol-lowed by the mode of grass cultivation in F. lacor forest and then poultry feeding in D. latiflorus forest. Fungi cultivation in E. robusta forest performed the least in terms of soil infiltration. Correlation analysis between soil infiltration and physiochemical properties showed significant positive correlation between soil infiltration and organic matter (β₁), moisture content (β₂), total soil porosity (β₃) and the 0.005 0.01 mm soil particle content (β₆) (P < 0.01). Also significant negative correlation existed between soil infiltration and soil bulk density (β₈) and 0.05 0.25 mm particle content (β₅) (P < 0.01). According to the correlation analysis, eight indexes of soil physiochemical properties (which significantly influenced soil infiltration) were noted. Comprehensive parameters of soil infiltration (a) and soil physiochemical properties (β) were obtained by the Principal Component Analysis. Then linear regression models of the indicators of soil infiltration and comprehensive parameters (a and β) were built as α= 0.249a₁+0.254a₂+0.252a₃+0.254a₄, β=0.167β₁+0.183β₂+0.200β₃+0.174β₄ 0.145β₅+0.131β₆+0.106β₇ 0.200β₈. In the models, a₁ was initial infiltration rate, a₂ was stable infiltration rate, a₃ was average infiltration rate, a₄ was total infiltration; β₄ was non-capillary porosity, β₇ was content of 0.001 0.005 mm soil particles.