The paper selected precipitation and geographic terrain data of 48rainfall gage stations in the Xin''anjiang River basin from 2000to 2010,and used inverse distance weight(IDW),radial basis function (RBF),ordinary Kriging (OK) and residual Kriging(RK) methods to interpolate and screened the best interpolation method and parameter according to the validated results. The results show that in annual scale the precision of RK method witch comprehensively considered the influences of elevation and latitude is the highest, then follows by IDW, OK and RBF; the search strategy of neighbor stations has greater impact on IDW interpolation accuracy; the precision of OK method witch considered anisotropic is higher, but the difference between the models is little. On monthly scale, the interpolation precision of IDW method to rich rainfall month is higher; the interpolation effect of RBF method to less and evenly distributed rainfall month is good; the data is more in accordance with normal distribution, the higher the interpolation accuracy of OK method; RK method is applicable to the month when the correlation between rainfall distribution and geographical factors is good.