Views diverge greatly on the relationship between actual evapotranspiration (AE) and potential evapotranspiration (PE). Penman showed that AE is determined in intensity by PE and changes proportionally as a function of PE. In contrast, Bouchet indicated that AE determines PE and varies inversely with PE. Based on nearly 30 years data from 432 weather stations and 512 hydrological stations in China, the two different theories, Penman's assumption and Bouchet's complementary relationship between AE and PE, were tested on nine river basins. With data integration technique, the complementary relationship between AE and PE was displayed entirely. A general model to estimate the actual evapotranspiration from nonsaturated surfaces by routine meteorological observations has been established on the basis of thorough analysis of the concept of PE. The results show that the calculations are all in error control of