In order to study the effects of different forest and grass restoration models on soil and water conservation on the hill slope in Karst trough-valley areas, the physical and chemical properties and infiltration characteristics of soils under different forest and grass patterns (forest land, pepper land and shrub grassland) on the consequent and reverse hill slopes in Karst trough-valley area were compared and analyzed by field investigation and indoor analysis. The results showed that:(1) The bulk density of forest land was the smallest, and the porosity, soil moisture content and organic matter content were the highest, followed by pepper land and shrub grassland. The soil organic matter content of the consequent hill slope was significantly higher than that of the reverse hill slope (P<0.05), but there was no significant difference in other physical and chemical indexes (P>0.05).(2) The infiltration curve of pepper land was steeper both on the consequent and reverse hill slopes, while that of forest land and shrub grassland were relatively slow. The infiltration rate of shrub grassland tended to be stable earlier (about 20 minutes), while the infiltration rate of pepper land and forest land began to decline slowly in about 30 minutes, and became gradually stable after 75 minutes. The soil infiltration capacity of forest and restoration modes on the consequent and reverse slopes was forest land > pepper land > shrub grassland(P<0.05). (3) In the infiltration model, Horton model had the best fitting effect (R²>0.712), compared with Kostiakov and Philip models. The determining coefficient R² of the regression equation between simulated stable infiltration rate and measured value was as high as 0.978 9.(4) The soil stable infiltration rate in the study area was positively correlated with total porosity, non-capillary porosity, saturated water content and pH value, but was negatively correlated with organic matter. Total porosity had the greatest impact on soil infiltration performance, while organic matter had the least impact. The multiple linear regression equation established by the above indexes could accurately predict the stable infiltration rate of soil. The results could provide a theoretical reference for the selection of vegetation restoration and the assessment of soil infiltration performance on the slopes of Karst trough-valley areas.