We improved a process‐oriented biogeochemical model of carbon and nitrogen cycling in grasslands and tested it against in situ measurements of biomass and CO2 and CH4 fluxes at five European grassland sites. The new version of the model (PASIM) calculates the growth and senescence of aboveground vegetation biomass accounting for sporadic removals when the grassland is cut and for continuous removals when it is grazed. Limitations induced by high leaf area index (LAI), soil water deficits and aging of leaves are also included. We added to this a simple empirical formulation to account for the detrimental impact on vegetation of trampling and excreta by grazing animals. Finally, a more realistic methane emission module than is currently used was introduced on the basis of the quality of the animals' diet. Evaluation of this improved version of PASIM is performed at (1) Laqueuille, France, on grassland continuously grazed by cattle with two plots of intensive and extensive grazing intensities, (2) Oensingen, Switzerland, on cut grassland with two fertilized and nonfertilized plots, and (3) Carlow, Ireland, on grassland that is both cut and grazed by cattle during the growing season. In addition, we compared the modeled animal CH4 emissions with in situ measurements on cattle for two grazing intensities at the grazed grassland site of Laqueuille. Altogether, when all improvements to the PASIM model are included, we found that the new parameterizations resulted into a better fit to the observed seasonal cycle of biomass and of measured CO2 and CH4 fluxes. However, the large uncertainties in measurements of biomass and LAI make simulation of biomass dynamics difficult to make. Also simulations for cut grassland are better than for grazed swards. This work paves the way for simulating greenhouse gas fluxes over grasslands in a spatially explicit manner, in order to quantify and understand the past, present and future role of grasslands in the greenhouse gas budget of the European continent.