This study aimed to compare the personality traits and attribution styles of runaway and normal girls. This is a comparative causal study. The population of the study involve 30 cases of runaway girls residing in the crisis intervention center of Nawab, affiliated to State Welfare of Tehran at autumn of 2015, they and were selected based on convenience. Normal girls who were matched counterpart for runaway girls are selected by clustering sampling and then systematic sampling from first and high school first and second degree. To assess personality traits, a NEOFFI.60 Personality Inventory short form was used and for attribution styles, Seligman and Peterson ASQ attribution style questionnaire was used. In order to analyze the data, we applied multivariate and univariate statistical variance analysis. The results of the study showed that there is difference between personality traits and attribution styles of runaway girls and normal girls that means in comparison to normal girls, runaway girls significantly suffer from a higher state of neurosis. Moreover, runaway girls tend to apply more pessimistic attribution styles in dealing with life events, in comparison with their normal counterpart girls, which means runaway girls attribute pleasant life events to external, unstable and specific factors, and attribute unpleasant life events to internal, stable and general. These indicate a cynical attribution styles. Therefore, understanding the personality and behavioral traits, as well as recognizing the type of adopted attribution style of runaway girls and normal girls in face of life events, is an effective step to prevent them from running away by identifying and screening individual girls at risk and providing treatment and adopting appropriate attribution style for these individuals.