Dairy farming is a significant contributor of agricultural greenhouse gas (GHG) emissions in Canada. Quantification of inter-farm variation in carbon footprint (CF) of milk (kg CO₂ eq kg⁻¹ fat and protein corrected milk, FPCM) and identification of management variables influencing this variation may be helpful in developing feasible strategies for reducing the CF of milk. This study estimated the cradle to farm-gate CF of milk production in 142 dairy farms across Ontario, a province that contributes about one third of total Canadian milk production. The objectives of the study were to: (i) quantify the inter-farm variation in CF of milk production using commercial farm data derived from an annual cost of production survey conducted by Dairy Farmers of Ontario and Canadian Dairy Commission, and (ii) identify key management attributes that significantly influence the CF of milk production. The analysis was conducted according to the methods recommended by International Dairy Federation. Results indicated that the CF of milk production on commercial dairy farms ranged from 0.44 to 1.73 kg CO₂ eq kg⁻¹ FPCM (mean ± standard deviation = 1.02 ± 0.24). The lower quartile of farms produced milk with mean CF of 0.77 ± 0.10 kg CO₂ eq kg⁻¹, while the quartile of high emitting farms produced with mean CF of 1.32 ± 0.18 kg CO₂ eq kg⁻¹. Low emitting farms produced milk more efficiently as indicated by significantly (P < 0.05) lower calving interval (months), age at first calving (months), whole farm feed-use (kg feed dry matter kg⁻¹ FPCM), land-use (m² kg⁻¹ FPCM), and direct energy-use (MJ kg⁻¹ FPCM) compared with high emitting farms. Results suggest that, at provincial scale, closing the productivity gap between low and high performing dairy farms may reduce the CF of milk production.