The article taking the representative natural vegetation (Populus euphratica) in the lower reaches of Tarim River as the test vegetation, the air temperature, net solar radiation, atmospheric humidity, wind speed on the canopy top edge, groundwater level and cross sectional area of the P. euphratica, as initial influence factors of model based on the field test. Multiple linear regression model, nonlinear regression model and complete quadratic regression model were established based on the least square method; transpiration of the populous euphratica in the lower reaches of Tarim River was simulated by using the model. Results showed that the three regression models showed better simulating effect. Simulation accuracy of the complete quadratic regression model was higher than that of the other two models. The air temperature, groundwater level and cross sectional areas of the P. euphratica were the main factors affecting the transpiration of populus euphratica forest in the models. Transpiration observed and simulated values of populus euphratica forest showed better correlation. The coefficients of correlation between observed and simulated for transpiration of P. euphratica of the three models were 0.6793, 0.7299 and 0.7574, respectively; the average relative error of simulated value with observed value for transpiration of that were 28.7%, 26.1% and 22.9%, respectively. All significance level of them passed α=0.05. In that the complete quadratic regression model test against experimental data showed a good agreement between observed and simulated transpiration of the populous euphratica, it presented the advantage of ease to use, impact factor determination easier and a certain precision. The model provided a new effective and feasible way for the evaluation of natural vegetation evapotranspiration and the calculation of ecological water requirements.