Though it has been shown recently that forest ecosystems affected by high rates of atmospheric N input are significant sources of N trace gases, reliable regional estimates of the source strength of such forests are missing. In this study, the biogeochemical model Photosynthesis and Evapotranspiration-Nitrification-Denitrification and Decomposition (PnET-N-DNDC), which simulates processes involved in N trace gas production and emission in forest soils on a daily scale, was used to calculate a regional inventory of N trace gas emissions from forest soils in southeast Germany for the year 1997. Prior to its use the model was further validated for climate and site sensitivity using multiyear observations of N trace gas fluxes at the Höglwald Forest site, Germany, and at forest sites in the Netherlands. On a regional scale the model estimates for Bavaria, Germany, that NO and N2O emissions from forest ecosystems in the year 1997 were ∼4.21 kt NO-N yr−1 and ∼6.64 kt N2O-N yr−1, respectively. Compared with total annual NOx emissions from combustion processes in Bavaria, total emissions of NO from forest soils are of minor importance [4.6%]. However, in summer, NO emissions from forest soils were of significant importance [8.8–22.0%] for the total regional NOx burden, since NO emissions showed a strong seasonal pattern with highest emissions during summer. Also, with regard to N2O, huge seasonal variations were found. Because of high N2O emissions during periods of freezing and thawing of forest soils, N2O fluxes in the winter period of 1997 also contributed significantly on the regional scale to total annual N2O emissions [∼38%]. Sensitivity analysis revealed that the accuracy of the N trace gas inventory strongly depended on the quality of regional input data, since the regional estimates of N trace gas fluxes calculated by PnET-N-DNDC were very sensitive to changes in soil texture, soil carbon content, site fertility, and, especially for NO emissions, to changes in soil pH.