通过遥感蒸散发模型估算实际蒸散发量的方法已被广泛使用，但精度提升仍是研究热点。作物生长模型在模拟作物蒸腾方面具有良好的机理性和精度。本文结合WOFOST作物生长模型和三温遥感蒸散发模型，提出了一种新的玉米农田遥感蒸散发估算方法。核心思路是本地化WOFOST模型，在验证其模拟精度后，利用其模拟的作物蒸腾数据，构造伴随率定函数，率定三温模型的蒸腾组分，然后合并率定后的土壤蒸发组分，得到玉米农田实际蒸散发估算值。以涡度相关系统观测的实际蒸散发量为参照，评估了该方法的估算精度和适用性。 结果表明，未经率定的三温模型蒸散发、作物蒸腾和土壤蒸发的相关系数分别为0.61、0.71、0.12，均方根误差为1.76 mm·d^-1、1.91 mm·d^-1、3.02 mm·d^-1，纳什效率系数均为负。仅率定土壤蒸发后，相关系数提高至0.77，但误差仍然较大（1.91 mm·d^-1），纳什效率系数为-0.74。利用WOFOST模拟的作物蒸腾率定三温模型后，估算值与实际观测的相关系数显著提高至0.89，均方根误差降至0.65 mm·d^-1，纳什效率系数达到0.79，表明该方法有效提高了三温遥感蒸散发模型的估算精度，并对其他遥感蒸散发模型的精度提升具有参考意义。

The method for estimating evapotranspiration using remote sensing evapotranspiration models has been widely applied, but there is need for research into improving its accuracy. Crop growth models exhibit strong mechanistic foundations and accuracy in simulating crop transpiration. This study integrated the WOFOST crop growth model with the three-temperature remote sensing evapotranspiration model to design a novel method for estimating remote sensing-based evapotranspiration in maize fields. The core approach involved localizing the WOFOST model, validating its simulation accuracy, and using its simulated crop transpiration data to construct an auxiliary calibration function. This function calibrated the transpiration component of the three-temperature model and combined it with the calibrated soil evaporation component to derive the evapotranspiration for the maize fields. Using actual evapotranspiration observed by an eddy covariance system as a reference, the estimation accuracy and applicability of the novel method were evaluated. The results showed that the correlation coefficients of evapotranspiration, crop transpiration, and soil evaporation in the uncalibrated three-temperature model were 0.61, 0.71, and 0.12, respectively, with root mean square errors (RMSE) of 1.76 mm·d^-1, 1.91 mm·d^-1, and 3.02 mm·d^-1, respectively, and negative Nash-Sutcliffe efficiency coefficients. After calibrating only the soil evaporation component, the correlation coefficients improved to 0.77, but the error remained large (1.91 mm·d^-1) with a Nash-Sutcliffe efficiency coefficient of -0.74. However, when the three-temperature model was calibrated using the WOFOST-simulated crop transpiration data, the correlation coefficient between the estimated and observed values significantly increased to 0.89, the RMSE decreased to 0.65 mm·d^-1, and the Nash-Sutcliffe efficiency coefficient reached 0.79. These results indicate that the proposed method effectively improves the estimation accuracy of the three-temperature remote sensing evapotranspiration model and offers insights for enhancing the accuracy of other remote sensing evapotranspiration models.