The mechanism by which residual straw incorporation affects nitrous oxide (N₂O) and carbon dioxide equivalent (CO₂-eq) emissions throughout the rice season under upland-paddy rotation systems is currently unknown. We aimed to elucidate its effect using a four-year experiment and meta-analysis in southwest China. In garlic–rice (GR) and wheat–rice (WR) systems, residual straw incorporation significantly decreased N₂O emissions (43.6% and 73.5%, respectively) and NO₃⁻-N concentrations, relative abundance of denitrifying bacteria (Anaeromyxobacter, Bacillus and Hyphomicrobium), and copy numbers of the norB and nosZ genes. Ultimately, the soil denitrification rate was reduced during rice tillering and full heading periods, but the soil organic nitrogen accumulation level was increased. The reduction in N₂O also resulted in an average reduction in the total CO₂-eq of the GR (23.4%) and WR (32.9%) systems in 2021–2022. In addition, the meta-analysis results showed that straw incorporation had a generally positive effect on soil N₂O emissions, but this effect was negative during the rice season in upland-paddy rotation systems, which supports the main results of our study. The path analysis results indicated that dry season residual straw incorporation slowed N₂O emissions during the rice season by increasing the soil C/N ratio and downregulating denitrifying microorganisms, thereby inhibiting the denitrification rate. Our findings challenge the understanding that straw incorporation increases greenhouse gas emissions during the rice season and suggest that future estimates of straw incorporation on methane (CH₄) emissions during the rice season should consider the offsetting effect of N₂O.