A new discriminant kernel width method is proposed.Based on this method,the application of KPCA and LPP based on discriminant kernel width method for palmpr int recognition is studied. Firstly,the within-class kernel width and the between-class kernel width are calculated by the training dataset and its label.The smaller kernel width is selected for the clusters of dense distri butions.On the contrary,the larger kernel width is selected for the clusters of sparse distributions.Secondly,KPC A and LPP based on discriminant kernel width method are used to extract the low dimensionality feature vectors. The cosine distance between two feature vectors is calculated to match palmprints.Finally,the new algorithm is tested on PolyU palmprint database. The experimental results show that compared with the traditional method,the rec ognition rate (RR) of the new algorithm is higher.The time for palmprint recognition is smaller than 0.6s.The effectiveness of the proposed method is verified.