The ¹⁵N natural abundance method has been widely used for evaluation of symbiotic N₂-fixation. The method inevitably requires a reference plant that reflects soil derived δ¹⁵N similar to that in a N₂-fixing target plant, for estimating the contribution of fixed N₂. However, it is often difficult to select a suitable reference plant. Recently, an alternative method was proposed using the difference in δ¹⁵N values between shoots and nodulated roots, which did not require a reference plant per se. Whether this method is applicable to a wide range of N₂-fixing plants having different growth habits and symbiosis types remains to be verified. To test the applicability of this method for perennial plants, we examined the difference in δ¹⁵N values between shoot and nodulated root (Δδ¹⁵N_s-nr), and that between shoot and root (Δδ¹⁵N_s-r) in 6-month-old plants grown in pots with different soil moisture regimes. The relationships between Δδ¹⁵N_s-nr and the percentage of N derived from atmospheric N₂ (%Ndfa) calculated from the conventional ¹⁵N natural abundance method, and between Δδ¹⁵N_s-r and %Ndfa were analyzed in N₂-fixing legume Lespedeza cuneata and N₂-fixing non-legume Elaeagnus pungens and Myrica rubra. A close correlation was found between Δδ¹⁵N_s-nr and %Ndfa as well as between Δδ¹⁵N_s-r and %Ndfa in Lespedeza cuneata, while no correlation was found in N₂-fixing non-legume species. The results indicated that Δδ¹⁵N signatures could be useful for estimating %Ndfa for N₂-fixing perennial legume (Lespedeza cuneata) in the first growth season but might not be applicable for N₂-fixing actinorhizal plants.