Until recently, IPCC emission factor methodology has been used exclusively to estimate GHG fluxes from soils for national inventories. However, the 2006 EPA greenhouse gas inventory includes estimates of N₂O and CO₂ fluxes from agricultural soils derived from simulations using DAYCENT and CENTURY. DAYCENT and CENTURY are process based terrestrial ecosystem models of intermediate complexity. Both models have been extensively tested with field data from various locations representing a wide range of vegetation types, soil classifications, and land management practices. DAYCENT simulates direct soil N₂O emissions, as well as indirect N₂O emissions from NO₃ leaching and N volatilization, for major crops and grazed lands at sub-county level resolution using crop area data from the National Agricultural Statistics Service (NASS). CENTURY simulates soil CO₂ fluxes for major crops and grazed lands using land management area information from the Natural Resources Inventory (NRI). IPCC emission factor methodology was used to estimate N₂O emissions and soil C fluxes for the minority (~10-15%) of agricultural land not simulated by the process based models. Meteorological data required to drive DAYCENT were acquired from DAYMET, while monthly weather required to run CENTURY are from PRISM. Soils data were acquired from STATSGO and NRI. Monte Carlo analyses were used to quantify uncertainty in N₂O emissions and soil C fluxes. Results are reported in the Inventory of U.S. Greenhouse Gas Emissions and Sinks.