In many developing countries, milk and meat produced from dual-purpose cattle systems contribute substantially to the national supply of animal-derived human-edible protein (HEP). Defining a main product and a co-product may be difficult in these diversified systems because of the large spatial variation and possibly rapid temporal variation in the contribution of each product within a farm and among farms. Few studies have investigated the impact of the milk-to-meat ratio output of cattle systems of developing countries on allocation of greenhouse gases when using Life Cycle Assessment (LCA). In this study, we analysed data collected from 677 Costa Rican farms with the objectives of 1) developing a field-based methodological framework to quantify the contribution of milk, meat, and HEP from milk and meat in dual-purpose systems relevant to developing countries, and 2) evaluating the effect of the functional unit (FU), co-products handling methods, and milk-to-meat ratio on how predicted enteric methane (CH4) would be allocated in an LCA of dual-purpose systems of contrasting orientation (towards beef vs. towards dairy). Farms were categorized based on farmer-reported farm income from milk sales relative to income from both milk and animal sales: Beef (BS; <1%), Dual-purpose towards beef (DP-B; 1–39%), Dual-purpose with dairy and beef in similar proportion (DP-DB; 40–59%), Dual-purpose towards dairy (DP-D; 60–89%), and Dairy (DS; 90–100%). Enteric CH4 emission was predicted from reported herd structure and emission factors per animal category (adult female, adult male, heifer and steer), and allocated to milk as the main product (FU = 1 kg of milk) with meat as a co-product or reported per kg of HEP. To allocate emissions to milk and meat, we used the following methods: (a) All-to-milk, (b) mass-based, (c) biological-based, and (d) system expansion. Milk and HEP production were significantly greater in DS compared to other systems. Emission attributed to milk varied from 66% to 100% when using all-to-milk, mass-based, and biological allocation methods but from −1% to 44% when using system expansion. Low productivity and high emission intensity of meat for the alternative system, leading to unrealistically high meat emission deduction, explained the substantially lower percentage of emission attributed to milk for system expansion compared to the other methods. All-to-milk and mass-based allocation methods were insensitive to change in milk-to-meat ratio of the system but increasing ratio altered substantially the partitioning of emission between milk and meat when using biological allocation and system expansion. Depending on the allocation method used, the imputed emission intensity of meat varied from zero (all-to-milk) to 13.4 kg carbon dioxide equivalent (CO2-eq) as CH4/ kg of meat (system expansion). Using HEP does not capture the importance of other nutrients in animal source foods but helped resolve LCA methodological issues of concern for dual-purpose cattle systems of developing countries.