In agricultural practices, reference evapotranspiration (ET0) parameters need to be specified for crop water requirement determination. Although, the crop water requirement is not an easy task. Numerous mathematical models and experimental methods have been developed for this purpose in the entire world. On the other hand, ten-daily and monthly average meteorological data are used in many of Irrigation and drainage projects which can affect the calculation of evapotranspiration in ten-day and monthly scale. In this paper, daily, ten-day, and monthly calculating periods were studied on accuracy of reference evapotranspiration estimation using FAO Penman Monteith, FAO Radiation, Modified Penman, Hargreaves, Priestley-Taylor and Makkink models in Sanandaj with cold semi-arid climate. Then, by selecting of the model with higher accuracy, the calibration was done. Moreover, to evaluate the accuracy of different models, lysimetric data were used from Sanandaj research station. The results showed that the modified Penman method can predict reference evapotranspiration with higher accuracy in all periods and calibration of the equation and has no effects on its accuracy. Also, the results showed that the computational periods have no significant effects on the accuracy of different models prediction.