The seven geographical factors, such as longitude, latitude, altitude, slope, aspect, regionalization (dryness) and vegetation, were usedas influencing factors of modern precipitation. Using R programming language, Arcgis spatial analysis and SPSS (Statistic Package for SocialScience) statistical analysis software, this paper discussed on how to select the influence factors and reconstruction method in spatial simulation ofmodern precipitation. Based on generalized additive model (GAM) by R language to realize the nonlinear analysis between modern precipitationand its influence factors, and the results were obtained that all the geographical factors are roughly linearly related to the mean annual precipitationexcept the factor of slope. The multivariate linear regression model was established between the geography factors and the multi-year averageprecipitation, and these factors were tested in multiple collinearity. The calculation shows that R2 of model was promoted in successive geographyfactors regression analysis, which means the type of geographical factors have a certain influence on the precipitation space simulation; the Pvalues of the slope orientation factor in nonlinear and linear models are larger than 0.05, indicating that the influence of slope orientation factor onthe precipitation is not significant; R-squared of linear regression model is 0.849 and deviance explanation of non-linear model is 89.6%, so asthat the results of the two models have a certain credibility.