Background and aims <p>Soybean is vital for global food security, necessitating understanding its response to environmental changes. We examined how growing season weather and rising atmospheric CO₂ affect seed traits of historical short-season soybean cultivars in Eastern Canada, with particular focus on water-use efficiency and nitrogen fixation dynamics.</p> Methods <p>Field trials data from 1993 to 2016 were analyzed for 14 cultivars spanning seven decades (1932–1992). Impacts of precipitation, mean maximum temperature (MTemp), mean maximum vapor pressure deficit (MVPD), and historical atmospheric CO₂ (which increased by 47&#xa0;ppm during the study period) on seed yield, protein and oil percentages, carbon isotope discrimination (Δ¹³C), and nitrogen isotopic composition (δ¹⁵N) were assessed using linear mixed-effect models and hierarchical partitioning analysis.</p> Results <p>Seed yield and Δ¹³C showed positive correlations with precipitation but negative correlations with MTemp and MVPD. Seed carbon percentage and Δ¹³C increased with atmospheric CO₂, while seed protein and oil percentages, and δ¹⁵N decreased. Hierarchical partitioning highlighted yield vulnerability during early reproductive stages (R1-R3, July) and protein yield sensitivity during pod-formation and seed-filling (R4-R6, August). Historical cultivar selection favored seed and oil yields, but not protein yield, Δ¹³C, and δ¹⁵N. MVPD emerged as a better predictor of seed traits than temperature.</p> Conclusions <p>Correlations between Δ¹³C, δ¹⁵N, and seed yield suggest selecting for higher yield may indirectly reduce water-use efficiency (indicated by higher Δ¹³C) and enhance biological nitrogen fixation (reflected by lower δ¹⁵N). These findings highlight the need to consider both seasonal weather variability and rising CO₂ in soybean breeding programs.</p>