Abstract: Based on the experimental data from the “863” National Key Project on water-saving agriculture, evapotranspiration of paddy rice under controlled irrigation condition was estimated by using a mended Penman-Monteith method presented in this paper with a variable canopy resistance. Results showed that canopy resistance could be estimated better by adopting Penman-Monteith model which modified by soil moisture content. Variation of canopy resistance under different water conditions was revealed at the same time. Relative error between measured crop evapotranspiration and evaluated one was reduced from 18.57% to 10.84% by using the mended Penman-Monteith model and the performance even was improved at growth stages when ground was not covered completely. The sensitivity of Penman-Monteith estimates of evapotranspiration with different input data and parameters was very dependent on the values of the soil moisture, air temperature and humidity. Regress parameters (c1, c0) showed less effect on crop evapotranspiration, which indicated the mended Penman-Monteith model would be available to estimate crop evapotranspiration under different irrigation conditions and at different areas when regress parameters were decided for one crop.