Abstract: Now a days, the number of websites is growing rapidly coupled with the hacking of the accounts which based on the traditional login method that depends on username and password. So, it is important to improve the security of these sites by developing a method of verifying the identity of users, we propose a face recognition system to achieve this target. In this study, we review a method to detect faces and eyes depending on skin color then depending on Viola-Jones algorithm. After that, we propose our system which captures a photo from web camera then detects faces, eyes and glasses by using the combination of (skin color and Viola-Jones) detection techniques. After face detection, we normalize the detected face photo and transform it into grayscale color then apply the special mask to delete the lower-right corner and the lower left corner. After that, we select facial features (depending on the face recognition algorithms) and configure a special template of this face. This template is stored in the special database in the case of registering a new user within the system or compared with pre-stored templates in the case of login to compute the similarity measurement using Euclidean distance measure from the input image. The result of our experiment reveals that the detection process of 110 images from the FERET database provides 100% accuracy in terms of face detection, 90% accuracy in terms of eyes detection and 69% accuracy in terms of glasses detection and the recognition process of 80 images (two images for each user) using 320 images from the ORL database related to 40 user (eight images for each user) for recognition algorithm training provides 93.75% accuracy in terms of using Eigenface, 92.5% accuracy in terms of using fisherface and 88.75% accuracy in terms of using LPP.