In this research, the spatial distribution of N, P and K in agricultural lands of Golestan province were evaluated using different geostatistical methods such as Kriging, Inverse Distance Weighting, Radial Basis Function, Inverse Multiquadric and Local Polynomial. For this porpose, 505 soil samples were collected from fields in 2008 and concentrations of these elements were measured for each sample. The performance criteria for evaluation were Mean Absolute Error, Mean Bias Error and Root Mean Square Error with the Jacknife methods. The results showed that Kriging was the best method for estimating N, P and K. Because, it showed the highest accuracy and the lowest error for estimating of these elements. Local Polynomial (degree3) was also found as the most unsuitable method for estimating of these elements. Semivariograms analysis indicated that N and K were the best fitted by Exponential models, also P by Spherical model. Result of zoning showed that amounts of N and P were lower in north and east regions of the study lands.