A novel Bayesian approach to face recognition based on wavelet transform is proposed. The original image is decomposed into low frequency and high frequency sub-band images by applying wavelet transform, the 2DPCA algorithm is used to compute the eigenvector space of the face. Firstly Bayesian approach is used to the low frequency sub-band, select the top 10 candidate images for matching in the post processing stage, secondly Bayesian recognition is parallel processed using these high frequency sub-band images. The face recognition result was gained through weigh-adding arraying. Its efficiency and superiority are clarified by comparative experiment on a subset of FERET face data.