The daily and total chamber measured N2O emissions from an intensively-irrigated pastoral clay loam-textured soil for the non-fertilizer, urea and urine treatments at Kyabram site in Australia were compared with the predictions by the DNDC, DAYCENT, WNMM and FASSET gas modules under WNMM simulation framework. The overall results suggested that these four gas modules were capable to estimate N2O emissions, but with different levels of success during different periods and at different time scales, based on the linear regression analysis and three types of correlation analysis compared with the continuous chamber measurements. The four gas modules had some difficulties in predicting the daily N2O emissions for the non-fertilizer treatment, but performed well in estimating the total amount of N2O emissions. For the urea and urine treatments, four gas modules gave satisfactory predictions of daily and total N2O emissions. Among them, the WNMM and FASSET gas modules demonstrated stable accuracy of predictions for N2O emissions from this urea/urine effected pastoral soil, and particularly the former was slightly better than the other. One artificial neural network was developed to test it performance in estimating daily N2O emissions for this study based on the WNMM-simulated soil water-filled pore space and mineral N contents of 0-20 cm topsoil and soil temperature at 10 cm, compared with the chamber measurements for the urea treatment. The simulation by the ANN was very promising, but it is still considered not capable to compete with the well-established process models due to its limitations for application. However, it has a very strong potential to be implemented for modifying the standard IPCC methodology for the N2O emissions inventory at large scales to account for the variations of climate, soil type, landuse, agricultural practices and topography.