Effective indicators of plant nitrogen (N) nutrition are needed to improve N management in grasslands. This is particularly the case for mixtures that rely on N fixation by legumes as a major N input, because no reference tool such as the nitrogen nutrition index (NNI) exists under these conditions. The aims of this study were to test the reliability of a plant-based index, the N concentration of upper leaves in the canopy (Nup), as a possible alternative for NNI in both pure and mixed grasslands. Data were gathered from four experiments covering a range of pure and mixed grasslands under different N fertilization levels. A cross-validation of Nup predictions against NNI in pure stands, and against two NNI-derived indices in mixtures, was performed. The Nup values appeared to be linearly related to NNI in pure stands of both grasses and legumes. The relationship was identical for the two groups of species and explained up to 86% of NNI variability. In mixtures, Nup also displayed a linear relationship with the two other tested indices, explaining 65% and 78% of variability. The conclusions of the three indices diverged with respect to strongly unbalanced mixtures, where the assumptions regarding the computation of NNI-derived indices were not met. Excluding these situations, the overall relationship between Nup and NNI proved to be identical for mixtures and pure stands. The results suggest that Nup is a valid criterion for plant N nutrition which applies to a broad range of grassland species and to mixture conditions.