Abstract: Using the simulation of AVIM GOALS(Atmosphere Vegetation Interaction Model and GlobalOcean Atmosphere Land System) which is developed in the Institute of Atmospheric Physics, Chinese Academy of Sciences(IAP/CAS), the sensible and latent heat fluxes are analyzed and compared with the ERA-40 reanalysis data, the results showed that AVIMGOALS basically reproduce main features of the annual mean climatologic state and seasonal cycle of the surface fluxes. The zonal distribution of the sensible heat flux is bimodal, and the latent heat flux has an obvious change from January to July, shifting from a single peak type to the double peak type. The spatial distribution shows that a high value area of the surface fluxes is mainly distributed in the Southern Hemisphere and low latitude regions of the Northern Hemisphere in January, where a big increasing of the surface fluxes occurs in the high latitudes of the Northern Hemisphere, and the surface fluxes in the Southern Hemisphere still remain greater values, and the spatial correlation coefficients of AVIMGOALS and ERA-40 reanalysis data are all above 0.01 significance level according to the t test. The standard deviation distribution of the surface fluxes is small in low latitudes and large in high latitudes, and it agrees with ERA-40 reanalysis data, which shows the good simulation performance of AVIMGOALS. The correlation coefficient of the annual variation in the Northern Hemisphere is larger than that in the Southern Hemisphere, which are 0.97 and 0.89 for the sensible heat flux and latent heat flux, respectively. Further analysis on the surface fluxes from AVIMGOALS and reanalysis data for annual changes of the correlation coefficient in spatial distribution shows that the larger correlation coefficient is mainly distributed in high latitudes of the Northern and Southern hemispheres. The most areas north of 30°N, South America, South Australia, and the Antarctic region are above 0.01 significance level according to the t test, which shows the coupled model captures the obvious signal of the seasonal change in these regions. In addition to the analysis of the correlation coefficient between the surface flux and its global mean value, the result shows that the correlation coefficient in most areas of the Northern Hemisphere is larger than 0.6, similar to the reanalysis data. The areas to the north of 20°N in the Northern Hemisphere and Africa beside 20°S are above 005 significance level according to the t test, which indicates the regions significantly contribute to the seasonal change of the global mean surface fluxes.