The aim is to study the influence of different vegetation types on CO2 fluxes at the same site, taking farmland as the object by using the simple biosphere model (SiB3) , select different vegetation types in Suzhou Dongshan site as input parameters of model to simulate CO2 fluxes, and compare with the observational data in Suzhou Dongshan site during 2011-04-16 to 2011-06-30. Results show that using corn as the site vegetation type input model to simulate CO2 fluxes showed a very good diurnal variation agreement compared to the measured data, but selecting the ordinary crop type input model to simulate CO2 fluxes, CO2 fluxes during the daytime in April and May were undervalued, and CO2 fluxes in June were overvalued; when we select the tea type to input the model, the simulated results significantly overestimated CO2 fluxes during the daytime in May and June. In addition, when using SiB3 model to simulate the daily CO2 fluxes, all three vegetation types input model can effectively simulate daily CO2 fluxes, but not significantly. It shows that choosing the right type vegetation input SiB3 model can effectively simulate the diurnal variation of CO2 fluxes, but can not effectively improve the daily CO2 fluxes simulation.