Background and aims Legume covercrops can partially offset synthetic N inputs by returning N-rich residues to soil for wheat production in arid regions. However, the outcomes of reducing N fertilizer remain underexplored across agronomic, nutritional, economic, and environmental dimensions. Methods This study was conducted from 2022 to 2024 in arid northwest China to evaluate the effects of rotating a mixed-legume cover crop (hairy vetch: common vetch = 1:5) with two consecutive spring wheat (Triticum aestivum L.) under six N management strategies: conventional spring wheat without cover crops receiving 225 kg N ha(-1) (TN100), and spring wheat after cover crop incorporation receiving 100% (CN100), 85% (CN85), 70% (CN70), 55% (CN55), and 0% (CN0) of the conventional N rate. Results Across the two wheat seasons, CN85 maintained grain yield and economic benefit comparable to TN100 (7,895 vs 7,451 kg ha(-1); 10,309 vs 11,045 CNY ha(-1)) and did not compromise grain quality, with 5.7% higher protein and similar wet gluten (similar to 32%) and sodium dodecyl sulfate sedimentation (similar to 17 mL). Total non-CO2 greenhouse gas emissions during spring wheat growing season (soil N2O and CH4 plus upstream emissions from agricultural inputs) were 9% and 17% lower under CN85 and CN70, respectively, compared to TN100. The CN100, CN85 and CN70 maintained water use efficiency (similar to 1.7 kg m(-3)) comparable to TN100, while CN85 and CN70 improved fertilizer-based N-use efficiencies by 24-36%. Entropy-TOPSIS identified CN85 as the optimal fertilizer strategy, with CN70 as a strong alternative. Conclusions Overall, moderate N reductions (15-30%) after legume cover crops incorporation maintained productivity and grain quality while reducing non-CO2 climate impacts.