It is well-known that relative growth rate (RGR) is closely related to C:N:P stoichiometry at the whole-plant level, yet it remains a misgiving to determine whether the growth-rate hypothesis is consistent between plant organs. Here, we examined RGR, C, N, P concentrations and their ratios of N:C, P:C, and N:P for four marsh herbaceous species (two gramineous species, Deyeuxia angustifolia and Glyceria spiculosa; two sedge species, Carex pseudocuraica and Carex lasiocarpa) grown in an increasing water level gradient (−5, 0, +5, and +15 cm relative to the soil surface) in the Sanjiang Plain of Northeast China. The applicability of the growth-rate hypothesis to leaf, stem, root, and total biomass was subsequently tested. With the increase of flooding stress, RGR and root mass ratio decreased to a certain degree for all species, whereas the above-ground biomass allocation was increased. The variation of N was much greater than that of P; hence the change in N:P ratios was determined mainly by N concentration. RGR was positively correlated with N concentration, N:C, and N:P for stem and total biomass when the data were pooled for all species but was negatively correlated with those for leaf and root organs. Furthermore, the theoretical predictions regarding the relationship between RGR and nutrient ratios were not always the case for each of the marsh herbs. Therefore, our results indicated that the organ-specific and species-specific for vascular plants should be carefully considered when using the theoretical association of growth rate with C:N:P stoichiometry.