As one of the major cultivation areas around the world, U.S. Corn Belt plays a vital role in the global food supply and agricultural ecosystem service. Here we use Ecosys model to simulate surface fluxes and maize/soybean yield over the agroecosystem in the U.S. Corn belt. We run Ecosys model at both site and regional scales. As for site scale, we selected 7 AmeriFlux sites located in the Corn Belt (i.e. Bo1, Br1, IB1, Ro1, Ne1, Ne2, and Ne3), and used observed NEE for model calibration and other variables (i.e. GPP, LAI, organ biomass, and ET) for model validation. As for regional scale, we run Ecosys at 0.25°×0.25° grid in the core area of U.S. Corn Belt (i.e. Illinois, Indiana, and Iowa) with the Princeton Surface Meteorological Forcing Dataset (PUMET) and Unified North America Soil Map (UNASM) as inputs. We resampled the observed county level corn and soybean yield from 2002 to 2014 into 0.25°×0.25° grids for model calibration at regional scale. To evaluate the performance of calibrated Ecosys at regional scale, we used the observed yield and the remotely sensed GPP and ET (from BESS) from 2015 to 2018 as benchmarks. Overall, with detailed configurations and rigorous calibration, Ecosys model can achieve good performance in simulating surface energy-water-carbon fluxes and crop yield at both site and regional scales in the US. Corn Belt.