Now-a-days Social Networking Sites (SNS) are increasingly gaining its importance to share thoughts, views, culture, status, image, video etc. in various domains such as academics, religion, politics, research and development. Apart of these strengths of SNS it has one major disadvantage which is inefficient authentication of user login. Due to this various types of fake message, personal or national threats, non-social activities, vulgar and harassing figures and videos etc. are posted by some imposters or non social personals. To the best of our knowledge there is no any technique reported in the literature to solve this very big social problem. To defeat this attack we have prepared our own SNS with embedding of face recognition technique at the time of registration of login in the site. Four holistic based algorithms PCA, LDA, ICA and SVM are used for evaluation and found very promising results ranges 79.77% ~ 93.10% at various protocols.