In order to investigate the feasibility of using rice critical nitrogen concentration as a nitrogen nutrition diagnosis index, a two-year positioning field gradient experiment using four rice varieties and four nitrogen levels (0, 75, 150, 225 kg·ha^–1 for early rice; 0, 90, 180, 270 kg·ha^–1 for late rice) was conducted for early and late rice. The critical dilution curves (N_c%) of the double-cropped rice based on leaf dry matter (LDM) were constructed and verified using the field data. Two critical nitrogen dilution curves and nitrogen nutrition indexes (NNI) of rice LDM were constructed for early rice [N_c% = 2.66LDM^−0.79, R² = 0.88, NNI ranged between 0.29–1.74, and the average normalized root mean square error (n-RMSE = 19.35%)] and late rice [N_c% = 7.46LDM^−1.42, R² = 0.91, NNI was between 0.55–1.53, and the average (n-RMSE = 15.14%)]. The relationship between NNI and relative yield was a quadratic polynomial equation and suggested that the optimum nitrogen application rate for early rice was sightly smaller than 150 kg·ha^–1, and that for late rice was about 180 kg·ha^-1. The developed critical nitrogen concentration dilution curves, based on leaf dry matter, were able to diagnose nitrogen nutrition in the double-cropped rice region.