Feed management decisions are crucial in mitigating greenhouse gas (GHG) and nitrogen (N) emissions from ruminant farming systems. However, assessing the downstream impact of diet on emissions in dairy production systems is complex, due to the multifunctional relationships between a variety of distinct but interconnected sources such as animals, housing, manure storage, and soil. Therefore, there is a need for an integral assessment of the direct and indirect GHG and N emissions that considers the underlying processes of carbon (C), N and their drivers within the system. Here we show the relevance of using a cascade of process-based (PB) models, such as Dutch Tier 3 and (Manure)-DNDC (Denitrification-Decomposition) models, for capturing the downstream influence of diet on whole-farm emissions in two contrasting case study dairy farms: a confinement system in Germany and a pasture-based system in New Zealand. Considerable variation was found in emissions on a per hectare and per head basis, and across different farm components and categories of animals. Moreover, the confinement system had a farm C emission of 1.01 kg CO2-eq kg−1 fat and protein corrected milk (FPCM), and a farm N emission of 0.0300 kg N kg−1 FPCM. In contrast, the pasture-based system had a lower farm C and N emission averaging 0.82 kg CO2-eq kg−1 FPCM and 0.006 kg N kg−1 FPCM, respectively over the 4-year period. The results demonstrate how inputs and outputs could be made compatible and exchangeable across the PB models for quantifying dietary effects on whole-farm GHG and N emissions.