Abstract: The Loess Plateau in China is one of the most severely eroded regions of the world. Since the implementation of "Grain for Green" ecological restoration project, biological soil crusts (biocrusts) were widely distributed in this region, which significantly affected surface runoff. Numerous studies have explored the effect of biocrusts on runoff. However, the conclusions were still widely different. In the Loess Plateau region, rainfall is mostly concentrated in June to September, and the rainfall duration is not fixed, which may affect the runoff characteristics of biocrustal slopes. This study investigated characteristics of runoff from biocrustal slope in different rainfall durations in the Loess Plateau region by using artificial simulated rainfall experiment. The experiment was conducted in the revegetated grassland of northern Shaanxi Provence, China. The experiment site was about 80 m×20 m, and the slope gradient was approximately 15°. The biocrust types were mainly moss crust and moss cyanobacteria mixed crust in this site and their average coverage was 79.2%. The dimensions of the experimental plots were 10 m×2.1 m (length×width).Canopy of higher plants in the plots was removed with scissors. According to the range of local biocrust coverage, two treatments were set: 1) slopes with undisturbed biocrust as a high coverage biocrust (the average biocrust coverage were 79.2%); 2) the slopes with removal of a part of the biocrusts by shovels, which simulated the low biocrust cover situation (the average biocrust coverage were 43.6%). Meanwhile, ploughing plots were set as the control group. The rainfall intensity was set as 90mm/h and the duration was1 hour. The results showed that the initial runoff time of biocrust slope was significantly reduced compared to the bare soil slope. The initial runoff yield time of bare soil was 1.59-3.04 times that of the biocrust slopes. There was a significant negative correlation between biocrust coverage and initial runoff generation time; Conclusion of the influence of biocrusts on runoff yield was contradictory during the first 15 min and 60 min. For 90 mm/h rainfall intensity, runoff from biocrust slope increased by 75.42% compared to bare soil when the rainfall duration was the 15 min. While, runoff from biocrust slope was decreased by 52.42% compared to the bare soil when the rainfall lasted to 60 min; the infiltration rate of soil moisture was affected by biocrusts. The infiltration rate of bare soil slope with 60 min rainfall was 34.30% lower than that with 15 min. The infiltration rate of high coverage biocrust slope with 60 min rainfall was only 6.38% lower than that with 15 min, which may cause the difference of runoff yield between bare soil slope and biocrust slope; the effect of biocrust on slope infiltration and runoff is closely related to rainfall duration. Different periods of rainfall are likely to lead to inconsistent conclusions. Therefore, the duration of rainfall experiment considering the factors of biocrust should be no less than 45 min. The study provides scientific evidences for explaining the differences in the effect of biocrusts on infiltration and runoff, and further clarifies the hydrological effect of biocrusts in arid and semi-arid areas.