Soil water repellency (SWR) is an important indicator of soil physical property, and closely correlated to the processes of soil erosion and hydropedological processes. It was conducted to understand the effects of biocrusts on SWR and its variation characteristics, as well as its influencing factors in the Loess Plateau of China. SWR was assessed by using water drop penetration time (WDPT) and water repellency index (Ri) method. Along the 292~595mm precipitation gradient, eight sampling sites were selected and the restoration age were similar at each survey sites. The SWR of biocrusts and the adjacent bare soil on each sampling site were measured by WDPT method and mini-disc infiltrometer, respectively. In the meantime, biocrusts thickness, moss biomass, moss density, bulk density, soil mechanical composition and soil organic matter were also measured to analyze its correlations with the SWR. The results showed that WDPT and Ri between biocrusts and bare soil were significant at each sampling sites (P<0.05). The WDPT and Ri of biocrusts was average increased by 54.85 and 5.80 times as compared with the corresponding soils without biocrusts. Whether it was on the biocrusts or bare soil, the WDPT and Ri showed a similar trend, which was decreasing from south to north along the precipitation gradient in the Loess Plateau of China. In addition, the spatial variability of biocrusts SWR was higher than that of bare soil. There was a positively relationship between the WDPT and Ri of biocrusts and biocrusts thickness, silt content, clay content, soil organic matter and annual average precipitation (P≤0.032). In addition, there was a negative correlation between the WDPT and Ri of the biocrusts and the biomass and sand content (P≤0.030). No-linear regression demonstrated that the WDPT and Ri of biocrusts could be simulated by biocrusts thickness and biomass. The simulated coefficients of determination (R2), root mean squared error (RMSE) and the Nash-Sutcliffe efficiency (NSE) of biocrusts WDPT was 0.97, 5.89s and 0.99, respectively. Correspondingly, the simulated result of biocrusts Ri was better than that of WDPT with the R2 of 0.97, RMSE of 1.10 and NSE of 1.00. In general, the WDPT and Ri of biocrusts was decreasing from south to north along the precipitation gradient on the Loess Plateau of China, and the SWR of soil surface can be increased by the biocrusts. The WDPT and Ri distribution of biocrusts was satisfactorily simulated by biocrusts thickness and biomass. The annual average precipitation could indirectly affect the biocrusts SWR by altering the crust thickness, biomass, and soil organic matter.