The type of vegetation present in an area directly impacts soil characteristics and has a significant influence on the formation and stability of soil aggregates. Water-stable aggregates are the best indicator of soil stability and reflect the ability of a soil to resist erosion on the Loess Plateau of China. Soil samples were collected from both forest and forest steppe vegetation types in the Dong Zigou and Zhang Jiahe catchments, which are located in the Yanhe Catchment of the An Sai Hilly-Gully Region. This study investigated the characteristics of soil water-stable aggregate using the Le Bissonnais (LB) and Yoder methods; the results of three LB treatments (fast wetting (FW), slow wetting (SW), and wetting stirring (WS)) were then compared and the mean weight diameter (MWD) of soil aggregates and erodibility factor (K) value were calculated. The comparison of the three LB treatments showed: 1) the WS treatment was the most destructive to the stability of soil aggregates and the soil water-stable aggregate particle size was mainly concentrated in the <0.2 mm size group; this illustrated that heavy rain and/or irrigation were the main factors destroying soil aggregates in the Loess hilly region; 2) the SW treatment was the least destructive to soil aggregate stability and soil water-stable aggregate particle size was mainly concentrated in the >2 mm size group; this shows that light rain or drip irrigation did not have strong damaging effects on soil aggregates; 3) the destructiveness of FW on soil aggregate structure was intermediate between that of the WS and SW treatments, indicating that the soil water-stable aggregate particle size was evenly distributed. The experiment shows that the primary destructive mechanism of soil aggregate structure in this region was dissipation and mechanical slaking. Aside from considering the impact of natural factors, one should use irrigation methods that cause minimal damage to soil aggregates as part of a larger effort to reduce soil erosion. The results of the LB method of treatment in general and the measurements resulting from the WS treatment specifically produced findings that were close to the results of traditional wet sieving (the Yoder method); the FW and WS treatments can simulate the effects of different amounts and intensities of rainfall and irrigation on the stability of soil aggregates. The basic principles of the Yoder method are reflected in the LB method. The LB method can not only simulate traditional wet sieving results, but can also explain the mechanisms involved in the disintegration of soil aggregates under different conditions in terms of the stability of soil structure; the LB method provides more comprehensive information than the Yoder method and can help determine the cause of the loss of soil structure. The soil aggregate structure can be accurately approximated using the LB method. The LB method can be appropriately applied to measure soil aggregate structure in the Loess Hilly-Gully Region. The MWD and the measure of the erodibility factor (K) are used to evaluate the water stability of soil aggregates. The results also showed that the MWD of soil water-stable aggregates in soils from the forest vegetation type was greater than that of soils from the forest steppe vegetation type, and SW > FW > WS; but the value of the erodibility factor (K) of soils from the forest vegetation type is lower than the value of K from soils of the forest steppe vegetation type. In areas of restored vegetation, soil water-stable aggregate varied from small particles into larger particles. Different vegetation types have diverse levels of soil organic matter content, diverse forms of soil aggregates and differences in the degree of soil aggregation present. These differences result in differences in soil erodibility and resistance to soil erosion.