To study the characteristics of soil moisture changes of yellow fluvo-aquic soils in the northern Anhui North Plain during different growing periods of maize, the Yanglou Watershed was selected as the study object, and correlation analysis of groundwater table, the number of days without rainfall in the previous period, evaporation from water surface and rainfall was conducted using 0～50 cm soil water content data from 2010-2018, and the relative contribution of different driving factors to soil water content were calculated using boosted regression tree algorithm. The results showed that the soil water content decreased with the increase of the soil depth throughout the maize growing season. Two typical soil layers could be classified according to the coefficient of variation, with 0～10 cm as the active layer and 10～50 cm as the less active layer. Soil water content showed a significant negative correlation with water surface evaporation, groundwater table and the number of days without rainfall in the previous period, but a significant positive correlation with the rainfall. The relative contribution of the number of days without rainfall in the previous period to soil water content was the largest, which accounted for 42.6%, followed by the groundwater table, which was 20.9%, and the crop growing period only contributed 7.3%, which was the smallest. This study can shed some light on the understanding of soil moisture changes, agricultural water management and water utilization.