Conversion of long-term pasture to cropping was investigated for its effects on nitrous oxide (N2O) emissions in a 2-year field experiment in the high-rainfall zone of south-western Victoria. Early termination (pasture terminated 6 months before sowing) followed by winter (ETw) and spring (ETs) crops and late termination (pasture terminated 1 month before sowing) followed by a winter crop (LTw) were compared with continuous, mown pasture (MP). Emissions of N2O were measured with an automated gas sampling and analysing system. Emissions from MP were the lowest throughout the study, resulting in annual losses of 0.13kg N2O-N ha–1 in the first and the second years of the experiment. N2O-N loss was 0.6kgha–1 from treatments without fallow in both years (LTw in 2013 and ETs in 2014). In the first year, annual losses from previous fallow in ETw and ETs plots were 7.1 and 3.6kg N2O-N ha–1, respectively. Higher annual N2O losses from treatments with fallow periods continued in the second year of the study and were 2.0 and 1.3kg N2O-N ha–1 from ETw and LTw treatments, respectively. High emissions were associated with N mineralisation and the accumulation of NO3-N in the soil during the extensive fallow period after early pasture termination or wheat harvest. Soil water content was a key factor influencing the temporal fluctuations in N2O emissions. Low emissions occurred when water-filled pore space was <30%, whereas high emissions occurred when it was >65%, suggesting that denitrification was the major source of N2O emission. Crop grain yield was not affected by the duration of fallow (and therefore timing of pasture termination) in the first year, but was lower (P<0.05) in the treatment without fallow in the second year. Terminating pasture late rather than early, thus reducing the length of the fallow period, is a practical way of reducing N2O emissions from mixed pasture–cropping systems.