Evapotranspiration is one of the most important factors in hydrological cycle. Potential evapotranspiration is used to design in irrigation and drainage networks and hydrological studies (Davis & Dukes, 2010). The conducted studies in different regions of the world have shown that the FAO-Penman-Monteith model can be presented as the most accurate method under various climatic conditions (Irmak et al, 2003, ASCE-EWRI, 2005, Jabloun & Sahli, 2008, Martinez & Thepadia, 2010 and Azhar & Perera, 2011). The calibration of the mentioned equation using lysimetric data in a wide range of climatic conditions and its applicability without the requirement for local calibration in different climatic conditions are the benefits of the FAOPenman-Monteith equation (Landeras et al, 2008). On the other hand, the Valiantzas model is one of the newest methods for estimating ETo. Advantages of using the Valiantzas equations in the estimation of reference evapotranspiration include simple application for spatial calibration, easy application for temporal distribution of reference evapotranspiration values, easy use for routine hydrological applications and simplicity of equations for other hydrological applications (a, b, c, Valiantzas, 2013). The purpose of this study was to evaluate and compare the accuracy of five different Valiantzas models for estimating reference evapotranspiration at the studied stations located in the northwest of Iran (Urmia Lake basin) including Urmia, Salmas, Mahabad, Takab, Tabriz, Sarab and Maragheh and providing the best version of the Valiantzas as the results of which are the highest concurrence with the FAO-Penman-Monteith method.