At present, using geostatistics for studying on these variables with random and structure is becoming one of the main spatial analytical methods within the field of environmental sciences. In hilly areas of Loess Plateau, due to the importance of soil moisture, many researchers have studied on its spatial variability and relationships between land use and topography using descriptive methods. However, studies on that the spatiotemporal variability of soil moisture using geostatistics are still inadequate in this area. In this paper, we use geostatistics combined with GIS to determine and quantify the spatial and temporal variability of soil moisture. The results indicated that soil moisture exhibited high fractal dimensions and clear spatial autocorrelation. For mean soil moisture of 0～50 cm, the seasonal changes of the sills were opposite to the mean soil moisture. During dry season, the spatial heterogeneity of soil moisture was high, and the sills showed high level. During wet and extremely conditionds, the sills decreased and were low. The ranges varying between 135 m and 160m were high during dry season and low during wet season. The nuggets, random variance, ranged from 0 2 to 1 5, and did not exhibit apparent seasonal pattern and loosely mimiced mean soil moisture. For profile soil moisture, the sills except for 5 cm increased with soil depth during whole observation period. The ranges varying between 60m and 160 m did not show the systemic change with depth. However, the ranges of soil moisture of 45 cm and 70 cm exhibited little fluctuation compared to 5 cm, 15 cm and 25 cm during study period. After a significant rain event, the changes of the sills and ranges with depth became more complicated. The nugget showed no clear pattern with depth, and tended to high value with high sill and low value with low sill. These suggest that the processes for influencing on profile soil moisture pattern were convoluted, and their relative roles were different during the observation period. The spatial dependence and seasonal variability of soil moisture were interpreted by land use, topography, precipitation and soil physical attributes. Their relative roles of influencing on soil moisture depended on different seasons.