Abstract: Festuca arundinacea as a kind of typical grass in northern China, it has the advantage of storing and infiltrating rain water. In order to research the runoff of F. arundinacea grassland, we studied the changing process of runoff coefficient under different coverage (bare land, 60%, 90%) and varied rainfall intensities (30, 60, 90mm/hour) in F. arundinacea grassland through artificial simulated rainfall experiments. And we also studied its relationship with different influencing factors. Taking bare land as control, we studied the runoff coefficient reduction rate of F. arundinacea grassland under varied experimental conditions. The results showed that: 1) The initial runoff time of F. arundinacea grassland decreased with the increase of rainfall intensity, and it prolonged with the increase of coverage. Simultaneously, the greater the rainfall intensity was, the smaller range the runoff coefficient reduction rate became. And the greater the coverage was, the more obvious the grassland delayed the initial runoff time. 2) The runoff coefficient of F. arundinacea grassland and bare land both rapidly increased and then tended to be stable with the increase of rainfall time. After simulation, it was found that the runoff coefficient all showed a logarithmic function with rainfall time. 3) Runoff coefficient of F.rundinacea grassland decreased with the increase of coverage, and it increased with the increase of rainfall intensity. Also, the lower the coverage was, the more significant the rainfall intensity affected runoff coefficient. But the larger the coverage was, the less significant the rainfall intensity affected runoff coefficient. 4) The ability of F. arundinacea grassland in reducing runoff coefficient was limited. With the increase of rainfall intensity, runoff coefficient reduction rates of two kinds of coverage grasslands were more close. When the rainfall intensities were 30, 60, 90mm/hour, the differences in runoff coefficient reduction rate between the 60% and 90% coverage grassland were 25%, 24%, 23%, respectively. 5) The regression equation of the runoff coefficient of F. arundinacea with rainfall intensity, coverage and rainfall time was established under experimental conditions. After verification, the equation could be applied to predict the runoff coefficient of F. arundinacea grassland under the experimental conditions.