In this research we studied soil water infiltration in twelve soil series (Saveh l, Saveh 2, Garetapeh, Dolatabad, Garechai, labar 1, labar 2, labar 3 Harisan 1, Harisan 2, Angilavand, Akbarabad)in the Saveh Plain from Markazi Province of Iran using a software program which can extract coefficients of models of water infiltration from observational data. The results indicated that in short time intervals the amounts of water infiltration in the Saveh l, Dolatabad, Labar 2, Labar 3 and Harisan 1, was best predicted by the Kastiakov model. The Saveh 2 , Garetapeh , Garechai , labar 1, Harisan 2 and Angilavand soils were best predicted by the Modified Kastiakov model and the Akbarabad series by the Philip model. In the intermediate intervals the Saveh l, Dolatabad and Akbarabad series were best predicted by the Kastiakov model and the Saveh 2, Garehtapeh, Garechai, labar1, labar 2, Harisan1, Harisan2 and Angilavand series were best predicted by the Modified Kastiakov model. The Philip model best predicted the Labar 3 series. In the longtime intervals in the Garechai, labar 2, Harisan 1 series the Kastiakov model best predicted the results while in the series of Saveh 1, Saveh 2, Garetapeh, Dolatabad, labar 1, labar 3, Harisan 2, Angilavand and Akbarabad series was best predicted by the Modified Kastiakov model. Infiltration rate estimate for all models had the highest fitness ratios to observed quantities. Accumulative infiltration values in the Philip model in the Saveh 2, Garetapeh, Dolatabad, labar 1, Harisan 2 and Angilavand series converted to accurate shape with passing time and had a descending trend.