Agriculture represents 14% of global anthropogenic greenhous gases (GHG) emissions, 46% of this amount being due to N2O emissions from soils (UNEP, 2012). N2O is a powerful GHG (IPCC, 2013) and its emissions from agricultural soils are related to physical-chemical parameters which depend on climate (temperature, rain…), soil properties (Robertson et al., 1989) and farming practices (irrigation, tillage, fertilization…) (Tellez-Rio et al., 2015). The IPCC Tier 1 emission factor remains widely used to estimate annual N2O budgets from agricultural soils by taking into account the annual amount of N input only. However, not taking into account the environmental controlling factors may introduce high uncertainty in N2O budget estimation. Our study aims at highlighting the key drivers of N2O emissions from two agricultural sites in the South West of France and at proposing an improved, simple and accessible methodology to estimate N2O budget at crop plot and seasonal scale. For this purpose, we benefited from a unique long time series of daily N2O fluxes (from 2011 to 2016) measured with 6 closed automated chambers on two ICOS sites with contrasted agricultural management (FR-Lam and FR-Aur). N2O annual budget vary from 1.04 to 7.96 kgN ha-1 yr-1 for winter wheat and maize crop, respectively. The effects of fertilization, rain and irrigation, plant development, spring mineralization and deep tillage on N2O emissions were investigated. Significant correlations between rain combined with fertilization and plant development, deep tillage or spring mineralisation was found with R² of 0.91, 0.99 and 0.85, respectively. We took advantage of these results to develop an empirical model, including N input quantity, residual N, leaf area index and water input in order to estimate seasonal and annual N2O budget. At the seasonal scale, the model output matched well with the observed budget, with a R² and a RMSE of 0.87 and 0.33 kgN ha-1 at FR-Lam and of 0.92 and 0.12 kgN ha-1 at FR-Aur, respectively. It also gave good statistical scores at the crop year scale with a R² of 0.96 and a low RMSE of 0.43 kgN ha-1 when binding data from both sites. Using the IPCC Tiers 1 methodology gave lower and more scattered results with a R² of 0.46 and a RMSE of 1.46 kgN ha-1. For sites where N2O fluxes are not monitored, that new methodology may be an alternative and a more precise methodology than the IPCC Tiers 1 approach. It has also the advantage to require only few and accessible input variables.