Abstract: Abstract: Aggregate stability (MWD) is an important indicator of soil structure stability, and closely correlated to the processes of slope hydrology and soil erosion. Up to date, few studies are conducted to investigate the variation of MWD and influencing factors on the regional scale. In order to study the variation characteristics of MWD and influencing factors in the Loess Plateau, a 508 km long transect including seven sampling sites along the precipitation gradient was laid out in this study. The MWD of cropland, grassland and woodland on each sampling site were measured, in the meantime soil mechanical composition, bulk density, soil organic matter, root mass density and vegetation coverage of each sampling site were also measured. The results indicated that there existed significant (P<0.05) difference between the MWD of cropland and the MWD of grassland and woodland. Cropland had the largest mean MWD, followed by woodland; the cropland had the lowest mean MWD. The ratios of the mean MWD of grassland and woodland to the mean MWD of cropland were 2.94 and 2.39, respectively. The reason why the mean MWD of cropland was lowest could be explained by the fact that the lower soil organic matter and root mass density of cropland, and the disturbance of MWD induced by the frequent tillage operations. The MWD of cropland showed a decreasing trend from south to north along the precipitation gradient. The changing trends of the MWD of grassland and woodland were similar, which could be attributed to the similarity of the near surface characteristics (such as non-disturbance, high plant residue return) of grassland and woodland. The MWD of the Yijun was slightly lower than the MWD of Fuxian, and then the MWD of sampling sites showed a generally decreasing trend from Fuxian to Erdos. The MWD of cropland, and grassland and woodland were all influenced by soil organic matter, annual average precipitation and annual average temperature. In addition, there existed a significant (P<0.05) relationship between the MWD of cropland and soil mechanical composition, however the MWD of grassland and woodland were closely related to root mass density. The poor relationship between the MWD of cropland and root mass density were attributed to its low root mass density and frequent tillage operation. The fact that grassland and woodland had high soil organic matter and root mass density which masked the influence of soil mechanical composition well explained that no significant (P>0.05) relationship was found between the MWD of grassland and woodland and soil mechanical composition. The annual average precipitation and annual average temperature influenced the MWD of cropland, grassland and woodland by affecting root mass density and soil organic matter. No-linear regression demonstrated the MWD of cropland could be satisfactorily simulated by median soil grain size and annual average precipitation with the coefficients of determination (R2) of 0.95 and the Nash-Sutcliffe efficiency (NSE) of 0.99. The MWD of grassland and woodland could be simulated by soil organic matter, root mass density and annual average temperature. The simulated result of grassland and woodland was not as good as that of the cropland with the coefficients of determination (R2) of 0.64 and the Nash-Sutcliffe efficiency (NSE) of 0.64, which under-predicted low measured MWD and over-predicted high measured MWD. The results of this study will be of great significance in understanding the spatial variation of aggregate stability (MWD) along the precipitation gradient and its influencing factors, and analyzing the regional change law of hydrological and erosional processes.