Fertilizer nitrogen (N) accounts for the majority of the greenhouse gas (GHG) emissions associated with intensive wheat production, and the form of fertilizer N affects these emissions. Differences in manufacturing emissions (as represented in the current carbon accounting methodologies) tend to favour urea, even when using the best available manufacturing technologies (BAT), whereas differences in fertilizer N efficiency and emissions of ammonia tend to favour ammonium nitrate (AN). To resolve these differences, data from 47 experiments in two large UK studies conducted from 1982 to 1987 and from 2003 to 2005 were reanalysed, showing that on average urea efficiency was 0·9 of AN (although mean ammonia emissions in 10 subsidiary experiments indicated an efficiency difference of 0·2); treating urea with a urease inhibitor (TU; AGROTAIN®, active ingredient N-(n-butyl) thiophosphoric triamide (n-BTPT)) brought its efficiency almost in line with AN; however, a significantly greater mean optimum N amount for TU (+0·1 of AN) was not fully explained. A standard response function relating wheat yield to applied AN was modified for degrees of relative inefficiency, thus enabling yields and GHG intensities (kg CO2e/tonne (t) grain) to be calculated using a PAS2050 compatible model for GHG emissions for any N amount of any N form. With AN manufactured by average European technology (AET), the estimated GHG intensity of wheat producing 8 t/ha was 451 kg/t; whereas with urea or TU made by AET it was 0.87–0.99 or 0.84–0.86 of this respectively. Using BAT for fertilizer manufacture, the grain's GHG intensity with AN and TU was 368 kg/t, but was 1·03–1·17 of this with untreated urea. The range of effects on GHG intensities arose mainly from remaining uncertainties in the inefficiencies of the N forms. Generally, economic margins and GHG intensities were not much affected by adjustments in N use for relative inefficiencies or different prices of urea-based fertilizers compared with AN. Overall, TU appeared to provide the best combination of economic performance and GHG intensity, unless the price for N as TU exceeded that for N as AN.