Abstract: By the ways of rainfall simulation, we studied the effects of different slope gradients andrainfall intensities on runoff and runoff coefficient of Cynodon dactylon lawn, and the dynamic laws of runoff coefficient during the whole runoff process . The results showed that: 1) runoff and runoff coefficient of the C. dactylon lawn increased with the increasing slope gradient and rainfall intensity, and the biggest runoff volume (33.77 L/m2) appeared under the condition of rainfall density 90mm/h and slope degree 20°. If keeping other conditions unchanged, the differences in runoff volume and coefficientat different slope gradients would decrease. 2) The initial runoff time was shortened with the increase of slope gradient and rainfall intensity; the initial runoff time occurred 60 min after rainfall at the density of 30 mm/h, while it was 30 min after rainfall at 60 mm/h and shortened to 20 min after rainfall at 90 mm/h. From the on set of runoff producing to the end of rainfall, the change of runoff coefficient followed a pattern of increasing rapidly at first and tending to be stable afterwards, which means that it increased at least 50% at first. Compared with other rainfall conditions, the coefficient at the 30mm/h rainfall density increased fastest, then tended to be steady after a certain period in the whole process; the coefficient increased as the slope degree and rainfall density increased, with a small difference within 0.05 when the slope degree or rainfall density were kept unchanged. 3) Based on grey correlation analysis, we come to the conclusion that the correlation degrees of rainfall intensity and slope gradient with runoff were 0.61 and 0.60 respectively, indicating that the rainfall intensity has more effect on runoff than slope gradient.The maximum correlation degree between rainfall intensity and the runoff was 0.84, while that between slope gradient and runoff was 0.91. 4) A linear equation between runoff, slope gradient and rainfall intensity was fitted by regression analysis, with the R2 value 0.89, suggesting a good fitting. The significant level of slope gradient and rainfall intensity with the total runoff volume was less than 0.01.