Accurately obtaining the interannual variation characteristics and spatial distribution of precipitation is crucial for agricultural production management，ecological protection and meteorological warning.Based on the annual precipitation data from 91 meteorological stations in Jiangxi Province from 2000 to 2018，spatial interpolation and crosschecking were carried out using multiple linear regression interpolation （MLR）considering elevation，inverse distance weighting （IDW） considering elevation，cooperative kriging （CK） considering elevation，and ANUSPLIN introducing elevation covariate to compare the four methods and cross-checked to compare the interpolation accuracy of the four methods.It is found that：（1） the mean annual precipitation in Jiangxi Province shows obvious longitudinal zonality，which is positively correlated with elevation.There is no significant correlation between latitude and precipitation.（2）The mean relative error（MRE），mean absolute error（MAE）and root mean square error（RMSE）of IDW and CK are significantly higher than those of ANUSPLIN and MLR，and the MAE and RMSE are about 40 mm and 80 mm more，respectively.（3） The interpolation accuracy is ordered as follows：ANUSPLIN ＞MLR ＞CK＞IDW；where the MRE of ANUSPLIN is 0.03%，MAE is 54.04 mm，and RMSE is 68.86 mm，which is the most accurate interpolation result.The study suggests that the ANUSPLIN method should be adopted as a priority for the study of precipitation interpolation in Jiangxi Province.