This paper presents a model to quantify total greenhouse gas (GHG) emissions from dairy farms. The model, which is based on a whole farm management approach, accounts for the variability that occurs in GHG emissions among farm production and management practices. The variation is accommodated in six dairy farm management systems (FMS), which broadly include typical dairy production systems in South Africa. These are pasture-based with high or low stocking rates, total mixed ration with high or low stocking rates, and partial mixed ration with high or low stocking rates. Three variations of functional units that were used to evaluate the environmental impacts of various FMS are defined as per animal unit = kg CO2-eq head-1 yr-1; per unit of farm area = kg CO2-eq ha-1 yr-1, and per unit of product = kg CO2-eq kg FPCM-1, where FPCM is fat and protein corrected milk. The results show a range of GHG emissions in CO2-eq among the FMS with various methodological approaches because of the large impact from different emission factors, which vary between accounting methods. The more detailed equations were recommended to effectively improve environmental impacts. These more detailed non-linear equations tended to predict more biologically realistic emissions when compared with the linear equations in which over or under-predictions of GHG were observed. The most prominent drivers for GHG emissions across all FMS were from enteric methane (CH4) and nitrous oxide (N2O) from soil management. Rankings among FMS varied according to output methodology and functional units. GHG emissions expressed per animal or per unit area differ greatly from those expressed from a given level of product. In conclusion, the accounting methodologies that are described in this paper to predict GHG emissions of animal-related origin performed sufficiently across all FMS, and could be applied to quantify the carbon footprint of dairy production systems in South Africa.