It may be good to estimate rice yield based on root bleeding snap weight (BSW) and leaf soil plant analysis development (SPAD) at grain filling stages in southeast China. There were seven treatments: (N-butyl) thiophosphoric triamide (NBPT) applied at rates of 0.00, 0.50, 0.75, 1.00, 1.25, 1.50% with urea @ 135 kg N ha -1 and CK (no NBPT or urea). All treatments received 75 kg P2O5 ha -1 (superphosphate, P2O5 12%) and 150 kg K2O ha -1 (potassium sulfate, K2O 60%), and then all fertilizer and inhibitor were applied once as base fertilizer. At grain filling stage, both of early and late rice, root BSW and leaf SPAD readings were obtained and grain yield determined at harvesting. In addition, the relationship between BSW, SPAD and grain yield was established. The results showed that the grain yield of rice, root BSW and leaf SPAD were increased significantly in application of NBPT with urea. Compared with 0% NBPT, grain yield, BSW and SPAD in 1.00% NBPT were improved by11.9-33.0%, 7.1-74.3% and 4.6-31.7% in early rice; and increased by 8.4-33.0%, 7.9-75.0% and 4.4-19.5% in late rice. However, these decreased when NBPT was less or more than 1.00%. On the other hand, the relationship between BSW, SPAD at grain filling stage and grain yield could be fitted by two-variable linear equation (p<0.05); and which provided better of validation experiment. Therefore, the grain yield could be predicted better by BSW and SPAD at grain filling stage in double rice cropping system.