Agriculture is the main source of nitrogen loading (EEA, 2012) and is the sector with the largest remaining emission reduction potential (Sutton et al., 2011). Furthermore, surpluses of nitrogen are forecast to grow in the next decade (FAO, 2008). The objective of this study is to evaluate the determinants of the use of nitrogen inputs in agriculture, and the effects of RDP implementation in Emilia-Romagna on preventing nitrate pollution through a spatial econometric regression model. Firstly, we carried out an estimation of both inorganic and organic nitrogen input in agriculture at the municipality scale for year 2000 and 2010..Secondly, we performed a Moran’s statistics and a LISA (Local Indicators of Spatial Association; Anselin, 1995) analysis in order test the data for local spatial autocorrelation. Finally, in order to provide a quantitative evaluation of the application of the agri-environmental measures on the impact of farming systems on water quality, we constructed two spatial regression models: INORGANIC AND ORGANIC. Spatial dependence was included to the regressions (OLS) through spatial lag and spatial error. The INORGANIC model explains more than 70% of the dependent variable and suggest that participation to the measure 214 is not likely to be important for explaining the reduction of the Inorganic Nitrogen in the municipalities of Emilia Romagna. Significant variables are farm’s size, population density, location in NVZs and share of certified organic surface on the UAA. The same regressors could not explain the dependent variable in the case of the ORGANIC model. The availability of better estimation of changes in nitrogen inputs, such as the calculation at the farm scale, would be an important component to allow for a more robust use of spatial econometrics in RDP evaluation related to Nitrogen reduction.