It is difficult to express the true soil moisture temporal and spatial variation in traditional methods, because it always reflects a certain spatial resolution or discrete temporal pattern, which is not a continuous process. In this paper, we use SWAT (Soil and Water Assessment Tool)model and multi-source remote sensing data to solve the problems caused by the effects of spatial and temporal scales of soil moisture monitoring. Explore the scientific regional scale and long-time-series soil moisture simulation method, and use the simulation results analysis soil moisture in different time and space in study area. The results showed the following:(1) soil moisture over the entire region was continued decline from 2008-2014. The major factors affecting this trend are temperature, precipitation, and human activities. (2) The distribution of soil moisture is affected by precipitation, temperature, and land cover. The distribution of soil moisture is characterized by decreasing from the mountainous area to the plains, and in the following order:woodland > agricultural land > grassland > sparse vegetation. (3) In the past 10 years, the area with low soil moisture shifted from the northern mountainous plains to the eastern part, southeastern plains, and southern mountains, one of the most obvious change was eastern part. (4) The variation in soil moisture in the four seasons was significant. Soil moisture in spring and autumn were affected by rainfall and temperature, and that in winter was affected by the snow melt and temperature. The sub-basin soil moisture change were consistent in same season in different years.