Direct measurement of soil moisture characteristics curve (SMC) due to spatial and temporal variation is labor and expensive. Therefore, estimation of SMC from basic soil properties which can be measured easily would be satisfactory. In this study, a dataset containing 50 particle size distributions (PSD) data of UNSODA were selected to verify parametric and nonparametric (ROSETTA model). Results indicated that SMC is easily predictable by means of PSD curve and parametric models could predict SMC more accurate than ROSETTA software. In order to determine the effect of the number of model input in prediction of SMC two methods were used, full PSD method using at least 4 mass particle frequencies, semi PSD method using sand, silt and clay percentages as a model input. Statistical analysis revealed that semi PSD method is the best fitted model to experimental data. The semi PSD method predicted SMC more accurately in comparison with other methods as a result of data independency. The predicted SMC is continuous and more reliable in drying. So the semi PSD method could be used in irrigation programming. Since, sand, silt and clay percentages are easily available soil properties and their spatial-temporal variability are approximately constant, our method can be used as an alternative to predict SMC in large scale studies.