The studies on revision of parameters of SCS-CN model have been done by lots of scholars in China, but most of them focused on using different methods to revise 1 of the 2 parameters. In order to investigate whether the simulated result obtained from revising the 2 parameters at the same time is better than only revising 1 parameter, and also to identify the applicability of SCS-CN model under 4 different kinds of land use, this paper chose Shixia small watershed, which is the upper area of Miyun Reservoir watershed, as the study area, and the monitoring data during 2006 to 2010 of Shixia small watershed was collected. 8 equal interval of initial abstraction ratio value from 0.05 to 0.4 was selected, and by using the arithmetic mean method, the curve number values of 8 different initial abstraction ratios were calculated respectively. This paper also evaluated the applicability of SCS-CN model in study area under 4 different kinds of land use by taking Nash efficiency coefficient as criterion. The results showed that: (1) In the study area, when the initial abstraction ratio values were 0.05 under 4 different types of land use conditions, the Nash efficiency coefficient were the highest, which were 0.75, 0.48, -1.11 and -0.24 under bare land, cropland, grassland and, respectively. (2) When we adopted SCS-CN model to predict runoff in upper area of Miyun reservoir, the recommend values of curve number under bare land, cropland, grass and woodland were 86, 73, 58, and 63, respectively. And the recommend values of initial abstraction ratio under 4 land use patterns were all 0.05. (3) The simulation results of the SCS-CN model could be improved to a certain extent by revising the 2 parameters at the same time. (4) SCS-CN model could be adopted in bare land and cropland in Shixia small watershed, but was not recommended to be adopted in grassland and woodland.