[Objective] The effects of different long-term vegetation restoration types on differences of soil surface infiltration and water storage in the loess gully region of Western Shanxi Province were studied in order to screen and enhance the ecological benefits of natural vegetation restoration as well as artificial vegetation restoration in Loess Plateau in the future, thereby providing a scientific reference for an in-depth understanding of the ecological benefits of vegetation restoration. [Methods] We measured infiltration in the top 30 cm of soil under four types of restored vegetation (Pinus tabulaeformis pure forest, Robinia pseudoacacia pure forest, Platycladus orientalis pure forest, natural forest). We also measured soil physical and chemical properties such as soil bulk density and mechanical composition, and monitored soil moisture dynamics in the four types of restored vegetation in the loess hilly-gully region. We calculated soil water storage and analyzed differences due to vegetation type, and conducted correlation analysis and principal component analysis. Three soil infiltration models were then fitted to the measured processes in order to compare their suitability for use in vegetation restoration areas of the loess hilly-gully region. Differences in soil infiltration and water storage between different vegetation restoration types were also determined. We also evaluated the ability of the models to supplement and improve the surface soil water infiltration pattern under different vegetation restoration types in the Loess Plateau region. [Results] ① Natural forest was more effective in improving the physical properties of the 0—30 cm surface soil compared with pure planted forests. Natural forest had greater water content and better water retention, and planted forests were more effective in improving the infiltration capacity of the soil in the study area compared with natural forest. ② The stable infiltration rate of soil under different vegetation types from largest to smallest followed the order of natural forest ＞ Robinia pseudoacacia pure forest ＞ Pinus tabulaeformis pure forest ＞ Platycladus orientalis pure forest (for the 0—10 cm layer); Pinus tabulaeformis pure forest ＞ Robinia pseudoacacia pure forest ＞ Platycladus orientalis pure forest ＞ natural forest (for the 10—20 cm layer); Pinus tabulaeformis pure forest ＞ Robinia pseudoacacia pure forest ＞ Platycladus orientalis pure forest ＞ natural forest (for the 20—30 cm layer); ③ The Horton model performed better than the Kostiakov and Philip models in fitting soil infiltration in the study area (R²=0.94), and will be helpful in better understanding the soil infiltration pattern in the study area. ④ The main physical characteristics affecting soil infiltration rate were soil bulk density, soil organic carbon, soil water storage capacity and soil clay particle content. [Conclusion]There were significant differences in soil infiltration properties due to different long-term vegetation restoration types. Compared with long-term replanted forests, the long-term natural forest had greater vegetation restoration benefits, and could effectively improve soil quality and enhance soil water retention capacity for the Loess Plateau.