Abstract: Overland flow sediment transport capacity is one of the important parameters of soil erosion model, and it is also a basis for accurate prediction of soil erosion. In this paper, the typical loess with two different median diameters (median size d50=0.095 mm, d50′=0.04 mm) in the sandy and coarse sand area was used in the simulation experiment of sediment transport capacity in the sink. The slope gradient range was from 7% to 38.4% and the unit discharge range was 0.000 14-0.005 26 m2/s. After unqualified analysis, the coupling relationship between the sediment transport capacity and slope gradient, unit discharge and the flow intensity indicators (the mean flow velocity, shear stress, unit stream power, stream power, effective stream power) were analyzed. The results showed that the sediment transport capacity exhibited an increasing trend with increased slope gradient and unit discharge, and the sediment transport capacity had a power function relation with the slope gradient and the unit discharge (R2=0.955). The index of slope gradient was 1.086 and the index of unit discharge was 1.372. So, the unit discharge had more significant impact on sediment transport capacity than slope gradient. Compared with Zhang’s, Wu’s, and Mahoodabadi’s models, Zhang’s model had a basic trial of this research , and the results of Bing Wu's model were all less than the measured values in this study, while Mahmoodabadi’s model was the opposite. The mean flow velocity could be predicted by a power function of slope and flow, the relation between sediment transport capacity and mean velocity was also a power function relation ,with an index of 1.9072 (R2=0.857 3, Nash-Sutcliffe coefficient NSE=0.879 5). Because the mean flow velocity was affected by many factors such as hydraulic parameters and surface conditions, the relationship between mean flow velocity and sediment transport capacity could not be optimized. For example, the index was significantly (P<0.05) affected by the median diameter of sediments. Shear stress could be used to predict the sediment transport capacity through power function relations and the index was 2.498 1 (R2=0.900, NSE=0.756 1), which was not closely related to the influence of soil particle viscosity. The stream power and effective stream power were better predictors than the shear stress, considering the critical stream power W0=36.5. The power function relation of steam power and sediment transport capacity was the best (R2=0.950, NSE=0.978), with an index of 0.920 8. Although the relationship between effective flow power and sediment transport capacity was not optimal, it was still a content of further research. The unit stream power prediction model got poor results for sediment transport capacity (R2=0.799 9, NSE =0.839 6), which was consistent with the results of Zhang’s, Bing Wu’s and others’ studies. So, the unit stream power could not be used as a flow intensity to predict the sediment transport capacity. Based on the dimensionless sediment transport capacity, the formula for calculating the flow capacity of the slope was presented. This study of the sediment transport capacity in the Loess Hilly Gully Area provided a new method for the soil erosion prediction model. It is of great scientific significance to predict slope erosion and to reveal the sediment transport mechanism of slope.