Carbon dioxide and nitrous oxide are two important greenhouse gases (GHG) released from cropping systems. Their emissions can vary substantially with climate, soil, and crop management. While different methods are available to account for GHG emissions in life cycle assessments (LCA) of crop production, there are no standard procedures. In this study, the objectives were: (i) to compare several methods of estimating CO₂ and N₂O emissions for a LCA of cropping systems and (ii) to estimate the relative contribution of soil GHG emissions to the overall global warming potential (GWP) using results from a field experiment located in Manitoba, Canada. The methods were: (A) measurements; (B) Tier I and (C) Tier II IPCC (Intergovernmental panel on Climate Change) methodology, (D) a simple carbon model combined with Intergovernmental Panel for Climate Change (IPCC) Tier II methodology for soil N₂O emissions, and (E) the DNDC (DeNitrification DeComposition) agroecosystem model. The estimated GWPs (−7.2–17 Mg CO₂eq ha⁻¹ y⁻¹; −80 to 600 kg CO₂eq GJ⁻¹ y⁻¹) were similar to previous results in North America and no statistical difference was found between GWP based on methods D and E and GWP based on observations. The five methods gave estimates of soil CO₂ emissions that were not statistically different from each other, whereas for N₂O emissions only DNDC estimates were similar to observations. Across crop types, all methods gave comparable CO₂ and N₂O emission estimates for perennial and legume crops, but only DNDC gave similar results with respect to observations for both annual and cereal crops. Whilst the results should be confirmed for other locations, the agroecosystem model and method D can be used, at certainly one selected site, in place of observations for estimating GHGs in agricultural LCA.