This study was tested a model for explain to social networks sites (SNS) usage with structural equation modeling (SEM). Using SEM on a sample of 475 high school students (35% male, 65% female) students, model was investigated the relationship between self-concealment, social appearance anxiety, loneliness on SNS such as Twitter and Facebook usage. It was hypothesized that SNS usage was predicted directly by self-concealment, was predicted indirectly by loneliness and social appearance anxiety. The SEM showed that SNS usage was significantly predicted directly by self-concealment and was significantly predicted indirectly by social appearance anxiety and loneliness. It was revealed that the fit indicies of the SEM created (?[superscipt 2] : 6.18, df: 2, ?[superscript 2]/df: 3.09, p = 0.04558, RMSEA = 0.065, GFI = 0.99, AGFI = 0.97, NFI = 0.98, NNFI = 0.95, CFI = 0.98, IFI = 0.98, RMR = 0.049; SRMR = 0.028) in fit values. Based on the results social appearance anxiety, self-concealment and loneliness of high school students are remarkable predictors in terms of SNS usage.