Assessing and improving the sustainability of dairy production systems requires an accurate quantification of greenhouse gas (GHG) emissions and major nutrient (N, C, P) flows associated with milk production at the animal, farm and field-scale. Life cycle inventory databases are, however, often based on rough estimates of GHG emissions and nutrient flows, and cannot account for spatially-explicit variation in these flows. Emission estimates can be improved when underlying processes influencing GHG emissions and nutrient balances are explicitly considered. We aim to improve life cycle inventory databases for milk production in the US by integrating process-based models into LCA data acquisition. We therefore perform a quantitative comparison of five process-based models to determine major nutrient flows and GHG emissions of milk production at the animal, farm and field-scale.