In order to evaluate the consistency of its intrinsic quality, tipping paper was scanned by NIR optical fiber diffuse reflection technology, the obtained spectrum was processed by Standard Normal Variation (SNV) and first differential, then the features were extracted by principal component analysis (PCA), and the calibration set was created according to Mahalanobis distance in principal component space to recognize the patterns of tipping paper made by different factories. The evaluation model was developed for the consistency of intrinsic quality of tipping paper made by the same factory. The results showed that: the model of calibration set could differentiate the tipping paper made by different factories effectively, and the evaluation model performed perfectly for evaluating the consistency of intrinsic quality of tipping paper made by the same factory.