ContextSoybean [Glycine max (L.) Merr.] is the most important legume grown worldwide. The effect of nitrogen (N) and sulfur (S) fertilization on seed yield is commonly studied in the United States (US). However, soybean yield response to fertilization remains inconsistent, partly due to the lack of standardized field designs and a better understanding of the plant nutrition processes underpinning yield formation. ObjectiveThe aims of this study were to assess the i) seed yield, (ii) plant N status (as N nutrition index, NNI), (iii) the contribution of N fixation, and (iv) the uncertainties on i), ii), and iii) in response to N-S fertilization using a uniform protocol across environments. MethodsTwenty-six trials in twelve US states tested five fertilization strategies that combined N and S at varying rates and timings. Using Bayesian statistics, seed yield response to fertilizer, NNI, and contribution of N fixation were analyzed at site and treatment levels providing both magnitude of responses and estimation of their uncertainties. From the significance of responses on seed yield, sites were split into two groups: non-responsive (18 sites) and responsive (8 sites). The NNI, ratio of the actual to the critical plant N concentration, was calculated to diagnose soybean N deficiency, and the N derived from the atmosphere (Ndfa, %) as N fixation contribution were investigated to better understand the source of plant N across all sites. ResultsRoughly for three-fourths of the sites, fertilization resulted in an unlikely (non-responsive) yield effect, with uncertainties ranging from 0.09 to 2.62 Mg ha−1. The other one-third of the sites were mainly responsive to S or both N + S, with the yield responses ranging from − 0.42–1.1 Mg ha−1 and uncertainties varying from 0.47 to 1.36 Mg ha−1. For the yield responsive sites, NNI presented a high proportion of deficiency (NNI<1) for most of the treatments, except for the “Full” signaling to a potential for yield response. Likewise, only 6% of the changes in Ndfa were not related to the treatment “Full”, and regardless of the seed yield response to fertilization, within the same site, soil and N fixation showed similar contributions to plant N demand. ConclusionDue to the high uncertainty in treatment response and contribution of N fixation, N fertilization is unlikely to increase yields, leading to non-profitable recommendations. Sulfur deficiency, on the other hand, should be explored under site-specific conditions. A decision support system should include appropriate diagnosis methods for identifying N and S deficiencies, such as NNI in soybean. Attainable maximum Ndfa did not appear to be affected by fertilization but largely varying depending on the site. ImplicationsFuture research should assess the role of soil and meteorological variables underpinning N fixation and soil N, along with the impact on seed quality composition, as a critical trait for this crop.