This research studies the spatial distribution quality of spatial interpolation methods. The research aims first to obtain unbiased estimator parameters based on regionalized variables in the study field. We used kriging techniques (called Local spatial interpolation) to rely on the variogram function with a fuzzy inference system, where fuzzy kriging is an extension of ordinary kriging. The second objective of this work is to estimate parameters of covariance models based on real data of soil chemicals, the data adopted in this research is taken from (100) real data for each soil chemical (Mg, Cl, and No3). These data are from Mosul Quadrangle in Mosul city in Iraq. After applying kriging techniques and a fuzzy inference system, we show the minimization of the estimation variance to choose the sentimental under uncertainty. We get the Smallest standard cross-validation of errors. Covariance models are described by exponential, and spherical model. With the best fitting models by the constraint of weights we note that the performance of the interpolation method is better by compared to the fuzzy system. In conclusion, the improvement does not rely on the statistical methods, but rather higher quality and large data of soil variables should be used to improve the prediction process. All programming computations are carried out in Matlab Language.