Ⅰ. Problem This study aimed to investigate the difference in problem behaviors of the middle-school students according to the results of BGT. In addition, this study also set out to examine the difference m the resulte of BGT by sexes of middle-school students. Ⅱ. Method 1.Subjects The subjects include 63 male and 60 female students attening the grade in Seoul. 2. Instruments The instruments used were the Bender-Gestalt test and child behaviors checklist measuring 9 problem behaviors in male(Somatic, Complaints, Schizoid, Uncommunicative, Immature, Obsessive-Compulsive, Hostile Withdrawal, Delinquent, Aggressive, Hyperative) and 8 problem behaviors in female(Anxious Obsessive, Somatic-Complaints, Schizoid, Depressed Withdrawal, Immature-Hyperactive, Delinquent, Aggressive, Cruel). These tests proved to have adequate reliability level. 3. Statistical Analysis The data collected was analyzed by the following method : t-test, one-way ANOVA, Duncans Multiple-range test, All the analyses were done by SPSS program for computer. Ⅲ. Results Major findings from this study are summarized as follow: 1. In male middle-school students, statistically significant differences were found among three in the case of 8 maturations levels of BGT. Problem behaviors except 1 problem behaviors(Delinquent). In female middle-school students, all the problem behaviors showed statistically significant differences among three maturation levels BGT. 2. There were not differences by sexes in the results of BGT. That is, Significant differences are not found between male middle-students' responses and female middle-students' responses. Male middle-school students' maturation levels are same female middle-school students' maturation levels.