Based on the available soil moisture (ASM) data collected from the 213 sampling sites over a small typical watershed in the hilly region of the Loess Plateau, spatial variability of soil available moisture at the scale of watershed, hillslope and gully and its seasonality (spring, summer and fall) was analyzed separately. Results show that the ASM, no matter at what scale, displays relatively strong spatial variability and ASM in gullies is higher in mean and spatial variability (Standard Deviation and Variation coefficient) than in hillslopes. The ASM also showed higher normality at gully and hillslope scales than at watershed scale. Spatial variability at all the three scales varies with the mean of ASM, and variation coefficient descends exponentially with the increasing mean. ASMs at the watershed and slope scales are closely and positively related to slope aspect, with correlation coefficient being higher than that its correlation with slope and elevation. However, the correlation coefficients of ASM with size of a gully and various landform factors are all low. The spatial variability of available soil moisture exhibits apparent seasonality. The ASM in fall, among the three seasons, is the highest in mean, but the lowest in variability. However, the ASM in summer is just the other way around. At the watershed scale, the ASM in summer is much higher than that in spring and fall in correlation coefficient with elevation, whereas it is in a reverse pattern at the slope scale. Besides, the uncertainty of the sampling of ASM and the estimation errors display a nonlinear descending trend with the increasing number of samples. However, when the number of sampling sites exceeds 20, the effect of increasing the number of samples becomes very limited. The findings presented here are expected to improve the understanding of seasonality of ASM spatial distribution of gully catchment in the Loess Plateau and could help design optimal sampling strategy of soil moisture.