Abstract: Taking Wanjinshan Township of Baoqing County as an example, the paper analyzes the spatial variability of four nutrient elements (soil organic matter, hydrolyzable nitrogen, rapidly available potassium, and available phosphorus) in ground substrate using ArcGIS 10.5 geostatistical analysis module, and studies the rationality of ordinary Kriging interpolation results for organic matter in different sampling densities. The cross validation is employed to assess interpolation accuracy by the root mean square error (RMSE) and correlation coefficient between predicted and actual values, ultimately determining the suitable number of sampling points. The results indicate that the spatial variability of all nutrient elements in ground substrate belongs to moderate variation, with the order of coefficient of variation (CV):available phosphorus > organic matter > hydrolyzable nitrogen > rapidly available potassium. The substrate effect of ground substrate nutrient elements all shows moderate spatial autocorrelation, among which organic matter has the largest substrate effect, while hydrolyzable nitrogen, rapidly available potassium, and available phosphorus have relatively smaller substrate effects. The order of autocorrelation distance of nutrient elements is hydrolyzable nitrogen > available phosphorus > rapidly available potassium > organic matter. The spatial variability analysis of ground substrate organic matter under different sampling densities shows that the suitable number of sampling points in Wanjinshan Township should be no less than 439, corresponding to a reasonable sampling density of more than 2 points per square kilometer (2 points/km²).