Abstract: The process of soil erosion is significantly impacted by rainfall-runoff pattern. To investigate the impact of runoff regimes on sediment yield and sediment flow behavior at slope scale, runoff and sediment data was collected and analyzed based on 65 individual runoff events from No.7 runoff plot at entire slope scale at Tuanshangou Catchment. Runoff process was characterized by runoff duration (T), runoff depth (H) and the ratio of peak discharge to mean discharge (flow variability, RPM) based on correlation analysis. Combined method of K-mean clustering, discriminant analysis as well as One-Way ANOVA was utilized to classify the runoff regimes. To quantify the relative impact of different runoff regimes on sediment yield from the same runoff amount (depth), 12 comparative groups were selected to conduct comparative analysis. Furthermore, dynamic indexes termed ξ were constructed with the method of multiple stepwise regression based on main indexes of runoff characteristics to depict the flow sediment behavior under different runoff regimes and runoff phases (rising limb and recession limb). The results showed that the runoff at entire slope scale could be classified into five regimes: Regime A with super-long duration, low flow variability, and minor total discharge (a particular regime); Regime B with relative long duration, medium flow variability, and large total discharge; Regime C with long duration, high flow variability, and large total discharge; Regime D with short duration, low flow variability, and minor total discharge of high frequency; Regime E with medium duration, medium flow variability, and medium total discharge. Area-specific sediment yield, mean suspended sediment concentration and maximum suspended sediment concentration showed great difference between different runoff regimes, which ranked in the order of C＞B＞E＞D＞A. This indicated that regime B, E, and C should be paid more attention to conduct runoff regulation. The difference of area-specific sediment yield between different regimes mainly derived from the variations in runoff amount (depth); and yet, the effect of altered flow-sediment relationship on area-specific sediment yield was masked. From another perspective, the difference of area-specific sediment yield originated from different regimes with the same runoff amount (depth) mainly derived from the variations in flow sediment behavior. Driven by the variations in sediment flow behavior, in comparison with regime A, the relative area-specific sediment yield from regime D, E, B and C were increased by 7.9 times, 6.3 times, 4.8 times and 4.5 times, respectively. In addition, the increase ratio decreased with the increase in runoff amount (depth). The optimum regression equation between suspended sediment concentration (S) and dynamic parameters (ξ) based on main runoff characteristics obeyed the function form of S = alnξ + b (R2＞0.5, sig＜0.001), which can interpret the main driving factors resulting in variations in sediment concentration under different runoff regimes and runoff phases. In conclusion, the results may provide some evidence for runoff pattern classification, construction of flow-sediment relationship, overall evaluation on the benefit of runoff regulation system at slope scale, as well as the further enrichment of the theory of slope runoff regulation.