Abstract: A Penman-Valiantzas model is developed based on the Valiantzas approach for estimating reference evapo-transpiration in China without wind speed data.Besides,the Hargreaves radiation equation is used for deriving a Pen-man-Valiantzas-Hargreaves model,in which only the relative humidity and the air temperature are required as the model input.The two new models are tested with the CLIMWAT database of UN Food and Agriculture Organization (FAO).A total of 156 Chinese stations are extracted from the database,which contain long-term monthly mean val-ues of climatic parameters.The result shows that the correlation coefficients between reference evapotranspirations cal-culated by the two models and the FAO Penman-Menteith model are 96.64% and 93.44%,respectively.The corre-sponding values of the root mean square error are 0.32 mm/d and 0.43 mm/d,respectively.These two new models can produce a more accurate reference evapotranspiration than that obtained by other two widely used models (Priest-ley-Taylor and Hargreaves).The Penman-Valiantzas model is further tested using the long-term measurement observed at the Yanqing Station in Beijing.The result of a 20-year monthly simulation reveals that the correlation coefficients between reference evapotranspirations calculated by the Penman-Valiantzas model and the FAO Penman-Menteith model as well as observed by the pan observation at the station are 97.84% and 86.38%,respectively.The daily simulation of reference evapotranspiration using the Penman-Valiantzas-Hargreaves model has also been improved. Our results demonstrate the applicability of the two new models in China in the absence of the wind speed and sun-shine hour data date.