Abstract – This paper aims to enhance the recognition rate in some Gabor face recognition techniques; the enhancement is done using two filtering stages. the first stage consists of applying adapted wavelet de-noising filter on the face database, this adaptive filter is composed of two filters (Bior 1.1 and Daubachies6) which are implemented successively both are at level 10 of decomposition. The second stage consists of computing and extracting the Gabor magnitude features using Gabor filter with 5 scaling and 8 orientations. All face images are loaded from the ORL database. As a result the Adaptive filtering technique produced good enhancement when applied on face recognition techniques, since two groups are used and are compared with our work. Each group includes four face recognition techniques: the first technique is Principal Component Analysis (PCA), the second technique is the Linear Discriminate Analysis (LDA), the third technique is the Kernel Principal Component Analysis (KPCA) and the fourth technique is the Kernel Fisher Analysis (KFA). The first group is applied with non-filtered techniques and the second group is filtered with Gabor filter, both groups are compared with the same group of techniques when de-noised with the adaptive double filter.