Recent papers have demonstrated the applicability of a calibrated bio-geochemical model – BIOME-BGC - for estimating current annual increments (CAIs) of Mediterranean forests. In this study, the above model is applied to assess the gross primary production (GPP) of nine beech forest sites in Italy using a previously produced data set of meteorological data over a ten-year period (1999- 2008). The obtained GPP estimates are integrated with relevant autotrophic respiration and allocation to obtain forest net primary production (NPP) averages for the same forests. The simulations are performed assuming different levels of ecosystem disequilibrium, i.e., progressively taking into account the effects of specific site history in terms of woody biomass removal and stand aging. The NPP estimates, converted into CAIs by means of specific coefficients, are validated through comparison with data derived from tree growth measurements. Results indicate that the modeling of quasi-equilibrium conditions tends to produce overestimated CAI values, particularly for not fully stocked, old stands. The inclusion of information on existing biomass leads to a partial improvement, while optimal results are obtained when information on ecosystem development phase is considered. The implications of using different NPP estimation methods are finally discussed in the perspective of assessing the forest carbon budget on a national basis. ; L'articolo è disponibile sul sito dell'editore http://www.sisef.it/