Agricultural soils have been determined to contribute at present about 60% to the global anthropogenic nitrous oxide (N₂O) emission due to N fertilization. It is a great challenge to guarantee high crop yields while reducing N₂O emissions under continuous application of nitrogen fertilizers. However, few field data sets are available for exploration of effects of fertilizer N regimes on soil N₂O emission in the Sichuan Basin, one of the regions the most intensive in agriculture in China. The main objectives of this research were to screen out optimal N fertilizer management practices which would not only reduce soil N₂O emission, but also maintain high crop yields. Static chamber-gas chromatographic techniques had been used to measure soil N₂O emissions from the purple soil under long-term N fertilization of pure synthetic N fertilizer (N), synthetic NPK fertilizer (NPK), pig manure (OM), pig manure plus synthetic NPK fertilizer (OMNPK) and incorporation of crop residues plus synthetic NPK fertilizer (ICRNPK). A plot with no fertilizer (NF) applied was set as control for emission coefficient calculation. In-situ field measurements were conducted through the wheat-maize rotation season from November 2012 to September 2013. N₂O emission showed a double-peak curve during the season regardless of fertilization regime. The peaks occurred in the first days after fertilization. However, the peak in the maize season was significantly higher than that in the wheat season (p < 0.05). Fertilization regimes influenced cumulative N₂O fluxes and grain yields significantly (p < 0.05). When the total N application rate was the same for all the treatments, 130 kg hm^-2 in the wheat and 150 kg hm^-2 in the maize season, Treatment N, OM, NPK, OMNPK and ICRNPK was 1.93, 1.96, 1.12, 1.50 and 0.79 kg hm^-2, respectively, in cumulative N2O flux, 0.62%, 0.63%, 0.33%, 0.47% and 0.21%, respectively, in emission coefficient, and 4.35, 11.95, 8.39, 9.77, 10.93 t hm^-2, respectively, in total annual grain yield. In comparison with Treatment NPK, Treatment OM enhanced N₂O emission significantly, whereas, Treatment ICRNPK significantly reduced N₂O emission and maintained high crop yields, as well. Therefore, incorporation of crop residues plus synthetic fertilization (ICRNPK) is recommended as an optimal fertilization regime in croplands of purple soil. Content of inorganic N (ammonium and nitrate) was the main controlling factor of soil N₂O emission. Therefore, the difference in N₂O emission flux was attributed to variation of the content of inorganic N induced by fertilization regime. That was also the reason why the fertilization regime, Treatment as ICRNPK, could effectively reduce N₂O emission. Soil water filled pore space (WFPS) might be another important factor affecting soil N₂O emission with a threshold effect. Therefore, it can be concluded that N₂O emissions from upland croplands under the wheat-maize rotation system can be reduced by optimizing timing and rate of N application in the light of crop growth and antecedent soil WFPS.