Abstract: Principal Component Analysis （PCA） has been proven to be an efficient method in pattern recognition.Recently，PCA has been extensively employed for face recognition algorithms，such as eigenface and fisherface.But PCA-based face-recognition systems are hard to scale up in applying because of incremental learning.This paper proposes a new Incremental Principal Component Analysis for Batch-Incremental data.First apply the space projection transformation based on the original PCs，then compute the unitary PCA in a lower-dimension transformed space.So the computation complexity can be reduced.And propose the kernel form of the incremental method at the same time.The experimental results on ORL face database verify that the presented approach is efficient.