Problem: Internet addiction has been emerged as a result of excessive internet misuse. In this study, analyzing the effects of depression, loneliness and self-esteem has been aimed in the prediction of the internet addiction levels of secondary education students. Method: The research is conducted according to the cross-sectional model as one of the survey models. The sample of the research is comprised of 292 students who continue their education in the first term of 2009-2010 academic year in Trabzon. Internet addiction, Beck depression, UCLA loneliness and Rosenberg self-esteem scale have been used as data collection tool in the research. Result: In consequence of the research a positive, mid-level and significant relation with internet addiction has come out when depression, loneliness and self-esteem variables are considered together. These variables explain 14 % of internet addiction's total variance. Conclusions: The relative order of importance of the variables on internet addiction is depression, loneliness and self-esteem. While depression and loneliness variables are significant predictors on internet addiction, self-esteem is not a significant predictor.