Empirical and process-based models simulating N2O fluxes from agricultural soils have the advantage that they can be applied at the scale at which mitigation measures can be designed and implemented. We compared bottom-up results from studies providing N2O fluxes at a regional/country or continental scale with estimates from the process-based model DNDC-EUROPE and from the TM5-4DVAR inverse modeling system. While the agreement between different bottom-up models is generally satisfying, only in a few cases a thorough validation of the result was done. Complex empirical or process-based models do not appear to have a better agreement with inverse model results in estimating N2O emissions from agricultural soils for countries or country-groups than simple ones. Both bottom-up and inverse models are limited by the density and quality of observations. Research needs to focus on developing tools that inherit the advantages of both methods.