Abstract: Potential evapotranspiration is an essential parameter for regional water balance analysis. To realize the accurate calculation of potential evapotranspiration when data are lacking, scientific evaluation of the applicability of the simplified potential evapotranspiration algorithm in the headwater area of the Yellow River is carried out. Based on daily observation data of temperature, precipitation, relative humidity, wind speed, and sunshine hours from 11 meteorological stations in the headwater area of the Yellow River from 1970 to 2018, the calculation accuracy of six simplified methods for potential evapotranspiration calculation are analyzed by using the FAO Penman-Monteith (PM) method as a reference method from the aspects of the year, month and spatial scales, including the Priestly-Taylor (PT) method, Doorenbos-Pruitt (DP) method, Hargreaves-Samani (HS) method, Rohwer (RO) method, Thornthwaite (TH) method and Blaney-Criddle (BC) method. The results show that the mean deviation between the HS method and the PM method, only 3.487 mm/mon, is the lowest in the headwater area of the Yellow River and the accuracy is the highest. However, the HS method does not consider the influence of the average relative humidity on potential evapotranspiration estimation, resulting in low accuracy in the humid climate of Hongyuan County, Henan County, Ruoergai County, Maqu County and Jiuzhi County in the southeast of the headwater area of the Yellow River. Therefore, this article introduces the average relative humidity factor to modify the HS method and evaluates the application effect of the improved HS method. It is found that after the HS method is revised, the mean deviation of the overall annual potential evapotranspiration in the headwater area of the Yellow River decreases from −22.008 mm/a to 6.174 mm/a, and the mean deviation of the monthly potential evapotranspiration decreases from 3.487 mm/mon to 1.031 mm/mon, respectively. On the spatial scale, the mean deviation of the above 5 counties decreases significantly, with an average drop of 5.33 mm/mon, indicating that the improved HS method can effectively solve the problem of low accuracy in the southeast of the headwater area of the Yellow River, and can provide a reference for the simplified calculation of potential evapotranspiration in the headwater area of the Yellow River.