No-tillage (NT), a practice that has been shown to increase carbon sequestration in soils, has resulted in contradictory effects on nitrous oxide (N2O) emissions. Moreover, it is not clear how mitigation practices for N2O emission reduction, such as applying nitrogen (N) fertilizer according to soil N reserves and matching the time of application to crop uptake, interact with NT practices. N2O fluxes from two management systems [conventional (CP), and best management practices: NT + reduced fertilizer (BMP)] applied to a corn (Zea mays L.), soybean (Glycine max L.), winter-wheat (Triticum aestivum L.) rotation in Ontario, Canada, were measured from January 2000 to April 2005, using a micrometeorological method. The superimposition of interannual variability of weather and management resulted in mean monthly N2O fluxes ranging from − 1.9 to 61.3 g N ha−1 day−1. Mean annual N2O emissions over the 5-year period decreased significantly by 0.79 from 2.19 kg N ha−1 for CP to 1.41 kg N ha−1 for BMP. Growing season (May–October) N2O emissions were reduced on average by 0.16 kg N ha−1 (20% of total reduction), and this decrease only occurred in the corn year of the rotation. Nongrowing season (November–April) emissions, comprised between 30% and 90% of the annual emissions, mostly due to increased N2O fluxes during soil thawing. These emissions were well correlated (r2= 0.90) to the accumulated degree-hours below 0 °C at 5 cm depth, a measure of duration and intensity of soil freezing. Soil management in BMP (NT) significantly reduced N2O emissions during thaw (80% of total reduction) by reducing soil freezing due to the insulating effects of the larger snow cover plus corn and wheat residue during winter. In conclusion, significant reductions in net greenhouse gas emissions can be obtained when NT is combined with a strategy that matches N application rate and timing to crop needs.