The critical nitrogen (N) concentration of a plant can be defined as the minimum nitrogen concentration required for maximum growth rate at any time. To determine the critical N concentration dilution curve for cotton, several field experiments with different levels of N application (0, 120, 240, 360, and 480 kg•hm^－2) were carried out in Nanjing and Anyang, standing for the ecological conditions in the middle lower reaches of Yangtze River Valley and Yellow River Valley in China, respectively. The results show that N concentration in shoot biomass declined with the growth stage after flowering. The relationship between the shoot dry matter and critical N concentration can be described by power equation put the equation here, with b=0.131 for both experimental sites, a=3.837 and 2.858 for Anyang and Nanjing, respectively. The results mentioned above support the viewpoint that the critical N concentration dilution curve for cotton is independent of ecological region. The maximum and the minimum N concentration dilution curves also follow a power equation put the equations here, with bmax=0.142 and bmin=0.158 for both experimental sites, and amax=3.530 and 3.208 and amin=3.055 and 2.251 for Anyang and Nanjing, respectively. The same estimate of b in critical dilution curve at the two experimental sites indicates that growth rate, density and pedoclimatic conditions do not affect the slope of the critical N dilution curve. The difference of the coefficients a between the two experimental sites shows that, the cotton plant in Anyang has a higher capacity of N accumulation in shoot biomass than in Nanjing for the same shoot dry matter. Due to its biological soundness, the critical N concentration dilution curve can be a theoretically sound and practically reliable tool for diagnosing the N nutrition status of plants. Based on the critical N concentration dilution model, the model of allometric relationships between crop N uptake at each N application level and accumulated dry matter in the shoot biomass, and the model of N nutrition index (NNI) were developed. The former can be used as an index for controlling of N application, and the latter can be used to express the N status of the cotton plants. If NNI=1, N nutrition is considered to be optimum, NNI>1 indicates excess N. NNI<1 indicates N deficiency. Based on the critical N concentration model, the model of N demand at different growth stages for potential growth and yield was developed. According to the allometric growth coefficient, NNI and N accumulation rate under critical N concentration, the following conclusion can be extracted: (1) Despite the difference of biomass and lint yield between Anyang and Nanjing, the eigenvalues of the dynamic biomass accumulation model were consistent. (2) The optimal rate of N application in Anyang should be higher than that in Nanjing, and the optimal N application rate is 360 kg hm^－2 and 240 kg hm^－2 in Anyang and Nanjing, respectively.Since the models developed in this study are based on the actual growth rate of the crop, it has the advantages over other models: it is crop specific, exact, simple and biologically sound. The models can be used directly to estimate the intrinsic crop nitrogen demand, and can be integrated, as a submodel, into the crop growth simulation models. The result of this paper has paved the way toward a timely precision nutrient fertilization.