150 near infrared spectrum of rice which obtained from southern China were collected. Combined with partial least square (PLS) regression method, the quantitative analysis models of rice moisture, protein, amylose, and gel consistency were established and verified by 30 predication sets. The coefficient of determination (R2) and the root-mean-square error of cross-validation (RMSECV) of the moisture, amylose, protein and gel consistency models were 0.9903 and 0.3728%,0.5603 and 1.4569%,0.9132 and 0.3054%, and 0.6780 and 5.0315%, respectively; standard error of prediction (RMSEP) were 0.3825%, 1.4650%, 0.5100% and 5.0521%, respectively. The results indicated that near-infrared analysis method can meet the demand of rapid determination.