Using data from eddy covariance measurements in a subtropical coniferous forest, a test and evaluation have been made for the model of Carbon Exchange in the Vegetation-Soil-Atmosphere (CEVSA) that simulates energy transfers and water, carbon and nitrogen cycles based on ecophysiological processes. In the present study, improvement was made in the model in calculating LAI, carbon allocation among plant organs, litter fall, decomposition and evapotranspiration. The simulated seasonal variations in carbon and water vapor flux were consistent with the measurements. The model explained 90% and 86% of the measured variations in evapotranspiration and soil water content. However, the modeled evapotranspiration and soil water content were lower than the measured systematically, because the model assumed that water was lost as runoff if it was beyond the soil saturation water content, but the soil at the flux site with abundant rainfall is often above water saturated. The improved model reproduced 79% and 88% of the measured variations in gross primary production (GPP) and ecosystem respiration (Re), but only 31% of the variations in measured net ecosystem exchange (NEP) despite the fact that the modeled annual NEP was close to the observation. The modeled NEP was generally lower in winter and higher in summer than the observations. The simulated responses of photosynthesis and respiration to water vapor deficit at high temperatures were different from measurements. The results suggested that the improved model underestimated ecosystem photosynthesis and respiration in extremely condition. The present study shows that CEVSA can simulate the seasonal pattern and magnitude of CO2 and water vapor fluxes, but further improvement in simulating photosynthesis and respiration at extreme temperatures and water deficit is required. Using data from eddy covariance measurements in a subtropical coniferous forest, a test and evaluation have been made for the model of Carbon Exchange in the Vegetation-Soil-Atmosphere (CEVSA) that simulates energy transfers and water, carbon and nitrogen cycles based on ecophysiological processes In the present study, improvement was made in the model in calculating LAI, carbon allocation among plant organs, litter fall, decomposition and evapotranspiration. The model explained in carbon and water vapor flux were consistent with the measurements. and 86% of the measured variations in evapotranspiration and soil water content. However, the modeled evapotranspiration and soil water content were lower than the measured systematically, because the model accounted that water was lost as runoff if it was beyond the soil saturation water content, but the soil at the flux site with abundant rainfall is often above water saturated. The improved model reprod uced 79% and 88% of the measured variations in gross primary production (GPP) and ecosystem respiration (Re), but only 31% of the variations in measured net ecosystem exchange (NEP) despite the fact that the modeled annual NEP was close to The observation. The modeled NEP was generally lower in winter and higher in summer than the observations. The simulated responses of photosynthesis and respiration to water vapor deficit at high temperatures were different from measurements. The results suggesting that the improved model underestimated ecosystem photosynthesis and respiration in extremely condition. The present study shows that CEVSA can simulate the seasonal pattern and magnitude of CO2 and water vapor fluxes, but further improvement in simulating photosynthesis and respiration at extreme temperatures and water deficit is required.