The objective of this study was to explain the response of cherry’s canopy transpiration(Tr) and conductance(gc) to different environmental conditions so as to evaluate the application feasibility of Jarvis model on cherry canopy scale modeling. Sapflow of 3 year-old potted cherry trees (Prunus avium L.) was continuously monitored using Granier sensors in Sijiqing orchard in Beijing from April to August, 2008. At the same time weather data and soil water content were also measured. Based on sapflow, gc was continuously simulated by back-calculated Penman-Monteith model. By analysis of cherry′s Tr and gc, multiple regression and Jarvis stomatal models were simulated and analyzed with cross-validation. The results showed that Tr had strong regularity and weak time lag, the gc had negative exponential relationships to vapor pressure deficit (VPD) under different radiation condition. The Various compositions of VPD, photosynthetic active radiation（PAR）and air temperature(Ta) were also used to obtain multiple regression and Jarvis models, which showed the precision was higher than that of multiple regression, the environment factors order rank affecting the simulation accuracy of model is VPD>PAR>Ta. TheⅡtypal Jarvis model using VPD and PAR was the best model for canopy conductance of cherry, with the lowest average relative error of only 12.12% and the lowest RMSE of 0.271.