Abstract: Evapotranspiration is a complex process and the only term that appears in both a water balance equation and a land surface energy balance equation. It is practically and scientifically significant for water resources management and ecological restoration to evaluate the applicability of reference crop evapotranspiration （ET₀） models. Based on the daily data of ten weather stations across a large climatic gradient in the headwater catchment of the Yellow River basin，China for recent 50 years，ET₀ was estimated with twelve different models，included FAO56 Penman-Monteith（P-M），Hargreaves-Samani，Turc，Irmak-Allen，Jensen-Haise，Priestley-Taylor，Makkink，Hamon，McCloud，Blaney-Criddle，Thornthwaite，and Holdridge methods. The first five methods are based on comprehension of temperature and radiation，and the Priestley-Taylor and Makkink are based on radiation，the last five methods are based on temperature. The climate data include maximum and minimum temperature and daily mean air temperature at 2 m height above the ground，wind speed，relative humidity，sunshine hours provided by the National Climate Center of the China Meteorological Administration. The western and southern of the study area is located in the Qinghai-Tibet Plateau；the eastern of the study area has some characteristics of the Loess Plateau，so an accurate estimate of reference crop evapotranspiration is the important work for the trend analysis of regional ecological changes and assessment of water resources. The P-M approach was recommended by FAO as the sole standard tool to calculate ET₀ and is a physically based technique and can be used globally without any need for additional adjustments of parameters. Taking P-M equation as the standard for evaluating the other methods，the linear regression and root mean squared error between P-M equation and the other simple equations were calculated，the results show that the best relationship and minimal difference were obtained with Priestley-Taylor or Makkink method，which is the best depending on location and season. The Makkink method is the best in alpine sites among growing season and the Priestley-Taylor method for other sites in all months and alpine sites in non-growing season. From the overall regional perspective and considering diversity of climate in study area，it is concluded that the Priestley-Taylor method is advisable to calculate ET₀ in case of only available-limited meteorological data. The methods based temperature in theory，such as Thornthwaite，McCloud，Blaney-Criddle and Holdridge，are poor-performed in daily ET₀ calculation and grossly underestimated ET₀；and its results are more consistent with P-M outcomes in the low elevation sites and the growing season than in the high altitude sites and the non-growing season，respectively. Some reports suggest that temperature-based ET₀ estimates perform as well as or better than physically-based ET₀ methods，but our results are contrary. This may be caused by that the methods driven by temperature do not reflect the climatic characteristics of bitter cold and intense radiation in high elevation region such as the study area.

Key words: reference evapotranspiration, calculation methods, applicability, Penman-Monteith equation, the headwater catchment of the Yellow River basin