Aims Rice cultivation plays a significant role in greenhouse gas emissions, particularly nitrous oxide (N2O) emissions. Although the effects of upland and flooded rice cultivation on soil N2O emissions have been reported, scholars have not comparatively investigated the mechanism underlying N2O emissions during the rice cultivation seasons of rice-based rotation systems. Methods Herein, a two-year field experiment including two rice cultivation modes, namely, conventional upland rice-rapeseed (UL-R) and flooded rice-rapeseed (FL-R) rotations, was conducted to determine the effect of different rice plantation models on soil N2O emissions in Central China. Non-rice treatments (UL-B and FL-B) during the rice season were also implemented to confirm the effect of rice plantation or soil condition on N2O emissions. Results Seasonal N2O emissions were higher in UL-R rotation than in FL-R rotation (1.54 +/- 0.16 vs. 0.71 +/- 0.20 and 2.57 +/- 0.28 vs. 0.76 +/- 0.04 kg N ha(-1) for the first and following rice cultivation seasons, respectively). Also, N2O emissions were higher in UL-B treatment than that in FL-B treatment during both rice seasons (2.45 +/- 0.07 vs. 1.43 +/- 0.35 and 3.74 +/- 0.37 vs. 1.16 +/- 0.08 kg N ha(-1), respectively). The yield-based N2O emissions were higher in the UL model than in the FL model (0.21 +/- 0.01 vs. 0.10 +/- 0.02 and 0.34 +/- 0.03 vs. 0.11 +/- 0.01 kg N ha(-1), respectively). The responses of N2O emission fluxes to soil ammonium (NH4+) and dissolved organic carbon (DOC) in UL rotation were stronger than those in FL rotation. Furthermore, total N2O emissions from non-rice treatments were higher than those from rice-cultivated treatments for both rice-based rotations. The increase in N2O emissions in UL-B treatment could be attributed to the higher abundance of amoA gene and elevated soil mineral nitrogen content compared to UL-R treatment. The higher amount of N2O generated in FL-B treatment than that in FL-R treatment was ascribed to the increased abundance of the nirS gene and the decreased abundance of the nosZ gene. The structural equation model supported that soil moisture, temperature, available C and N, and ammonium oxidation-related functional genes explained more than 70% of the effect on soil N2O emissions in UL rotation. Meanwhile, soil moisture, temperature, available N, and denitrification-related functional genes explained 80% of the effect in FL rotation. Our results highlight that the response of N2O emissions mainly relies on soil available N and C and critical functional genes, which may contribute to improve the understanding for N2O mitigation strategies. Conclusions These findings highlight the importance of rice plantation and their contribution to decreased field N2O emission, and suggest that soil available C, N, and critical functional genes should be considered when investigating N2O mitigation pathways during rice cultivation seasons.