Abstarct: Soil is considered as one of the main factors in the production of agricultural products, and its use must be baxsed on sound and scientific principles to be used in the production of products as a sustainable source. Any mistakes in exploiting it will result in the loss of this valuable resource. Today, the development of a more precise method for agricultural production is one of the ways to minimize the pressure from agricultural operations on the quality of the environment. In recent years, the use of methods such as artificial neural network and backup vector machine have been considered for soil characteristics estimation. Therefore, in order to estimate the nitrogen and sulfur content of the soil, using artificial neural network model and support vector machine model, the pH, EC, texture and organic carbon properties were used and the results were compared together with the multivariate linear regression model . For estimating and analyzing the parameters, the coefficient of determination (R2), mean square root error (RMSE) and mean absolute error (MAE) were used. The results of this study showed a higher efficiency of support vector machine model with R2 (0.99, 0.99), RMSE (0.0074, 1) and MAE (0.0057, 0 /77318) for N test and Soil sulfur is compared to artificial neural network model and linear multivariate regression model.