Picea crassifolia is the main afforestation tree species in the Loess Plateau in China. Vegetation reconstruction and forest stand regulation can be guided reasonably by studying water consumption characteristics in this area. It is important for forest stand to strengthen stability and to improve water use efficiency (WUE). In order to reveal the canopy transpiration characteristics of Picea crassifolia during the growing season and the response of the overall canopy stomatal resistance to environmental factors, and to understand the equation of Penman-Monteith whether or not suitable for scale of canopy of Picea crassifolia, we used thermal dissipation probe (TDP) to measure sap flow of Picea crassifolia and monitored the factor of environment by weather station (DAVIS). Then we framed the regression model between Canopy overall stomatal resistance (r_sT) and the environmental factors by inverse method. Lastly, we used equation of Penman-Monteith to predict the transpiration of Picea crassifolia during the growing season. The Root Mean Squared Error, Mean Absolute Deviation and Mean Relative Error would be the indexes to test and verify the accuracy of model we framed when we compared the predicted value with the measured value. The results show that (1) the daily evapotranspiration of Picea crassifolia increases firstly and then decreases with the monthly changes during the growth season. The ratio of monthly evapotranspiration to potential evapotranspiration is Jul (79.68%) > Aug (72.71%)> Jun (72.53%) > May (67.08%) > Sep (66.48%) > Oct (64.29%). (2) There is an obvious Lag-time phenomenon between sap flow and meteorological factors about 0.5 h. (3) The r_sT is positively correlated with relative humidity (RH), and negatively correlated with atmospheric temperature (T) and saturated water vapor pressure difference (VPD). (4) We applied the established multi-factor regression model and Penman-Monteith equation to simulate and verify transpiration of Picea crassifolia. The results indicated that the cumulatively average relative error is 14.381%, the average absolute error is 0.160 mm, and the root mean square error is 0.2. The simulation effect is basically feasible. In summary, the equation of Penman-Monteith has good applicability on the stand canopy scale. We can accord to the established multi-factor regression model combined with the equation of Penman-Monteith, three meteorological factors of saturated water VPD, T and RH can be used to simulate the evapotranspiration on any day, avoiding the inconvenience of long-term observation.