Abstarct: Knowing the hydraulics properties of the soil, such as the moisture reteention curve, is a prerequisite for modeling water movement and transfer of salts in the soil. Direct method of estimating this feature is costly and time consuming. In this study, moisture retention curve was Obtained 80 soil samples collected from agricultural fields of Shahrood (Bastam). For this purpose, parametric and point pedot transfer functions were developed baxsed on multiple linear regression and artificial neural network. Physical parameters (such as bulk density, particle distribution) and chemical (like organic matter and lime) were measured as inputs of the models in the laboratory. The soil moisture content was measured at the potential of 0, 10, 30, 50, 100, 300, 500, 1000 and 1500 kPa by means of pressure plates. Characteristics of 56 soil samples were considered for developing of models and characteristics of 24 soil samples for validation of the results. The results showed that the Derner model by assigning values of RMSE =0/0069, GMER =0/994 at the agricultural capacity point and values of RMSE =0/012, GMER =1/02 at the point of permanent wilting, the Van Genaghan model with assigning values of RMSE = 0/0086, GMER =0/976 at the crop capacity point and values of 0RMSE =0/024, GMER =1/08 at the wilting point had the best results and the model of the Campbell by assigning the values of RMSE = 0/0003, GMER =0/228 had the weakest performance in the estimation of crop capacity moisture. The results of comparison of MLR and ANN methods in moisture point estimation showed that MLR model with assignment of values RMSE = 0/087 and RMSE = 1/166 in suction 0.3 and 15 times and ANN model with assignment of RMSE values = 0/076 and RMSE = 0/086, respectively, in suction 0.3 and 15 times with the worst and best performance.