Aiming to clarify the mechanism and prediction of CO2 exchange between a plant community and the atmosphere, a multilayer model of CO2 exchange having a feedback function is proposed. This model can calculate not only the profile of CO2 flux within and above a canopy, but also the profile of meteorological elements such as solar radiation, long-wave radiation, air temperature, specific humidity, wind velocity and CO2 concentration. The profiles are calculated using information on the canopy structure, leaf characteristics, and meteorological elements observed over the plant community. In order to validate the reproducibility of CO2 flux over the plant community by this model, we observed CO2, latent heat and sensible heat flux over a planted forest, albedo and the profile of relative wind velocity, downward diffuse solar radiation and CO2 concentration within the canopy. By applying this model, we found that it could successfully reproduce not only CO2 flux, but also latent heat and sensible heat flux, over the canopy and reproduce the profile of these meterological elements. Moreover, numerical experiments on CO2 flux were carried out using this model to investigate the influence of the characteristics of canopy structure on CO2 flux. As a result, we found the followings: (1) When the values of LAI are 6 and 8, the fluctuations of CO2 fluxes over a canopy are more sensitive to the meteorological elements over the canopy than when the values are 1 and 3. (2) CO2 absorption by a community is not always greater during high solar elevation, as the average value of leaf inclination is greater. (3) Among the three canopies studies in the numerical experiments, one whose distribution of leaf area density is almost even absorbs CO2 gas the most.