Improving nitrogen use efficiency in crops is of great importance both in crop production and environmental protection. Genetically improving crop varieties so as to make full use of the potentials in nitrogen uptake and utilization has been considered as one of the most economical and effective ways to enhance nitrogen use efficiency in crops. Nitrogen use efficiency depends on the ability of nitrogen uptake and utilization in rice plant. Many studies have shown that there exist genotypic differences in nitrogen use efficiency for biomass and grain production in rice. Yet, little information in the characteristics of yield components, biomass production and allocation, nitrogen uptake and allocation, characteristics of source and sink, and structure of root system in rice varieties with higher nitrogen use efficiency has been reported. In this study, 88 and 122 conventional indica rice cultivars were solution-cultured in 2001 and 2002, respectively. Leaf area index (LAI), sink potential, dry matter weight and nitrogen content of different organs were measured. The tested rice cultivars were clustered into 6 types (i.e. A, B, C, D, E and F,A was lowest, F was highest) based on their nitrogen use efficiency for grain output (NUEg) level by the MinSSw method, to study the difference of indexes of source and sink and find out the key one from them in conventional indica rice with different types of NUEg. Results showed that the difference of NUEg in conventional indica rice cultivars used in this study was very large, the average NUEg of A, B, C, D, E and F were 20.51, 31.04, 35.64, 39.46, 43.55 and 50.92 g grain g-1N in 2001, and 24.33, 31.61, 35.83, 39.06, 43.51, 50.00 g grain g-1N in 2002, respectively; the basic characteristics of conventional indica rice cultivars with higher NUEg were that LAI at heading stage was small and decrease of LAI during the grain-filling stage was slow, while net assimilation rate was high, sink potential was large, sink potential per unit leaf area , sink potential per unit dry matter weight and sink potential per unit N at heading stage were high. Multiple stepwise regression analysis showed that NUEg was significantly influenced by sink potential per unit N, sink potential per unit dry matter weight, grain yield per unit sink potential, LAI at heading stage and decrease of LAI from heading to mature stage, and the determination coefficient ranged from 0.797–0.854. Path analysis showed that effects of sink potential per unit N, sink potential per unit dry matter weight and grain yield per unit sink potential on NUEg were larger than those of LAI at heading stage and decrease of LAI during the grain-filling stage.