This study was conducted to construct a hypothetical model of depression in Korean adolescent women and validate the fit of the model to the empirical data. The data were collected from 345 high school girls in Seoul, from May 1 to June 30, 1998. The instruments were the Body Mass Index, Physical Satisfaction Scale, Family Adaptatibility and Cohesion Evaluation Scale III, Family Satisfaction Scale, CES-D and School Adptation Scale. The data were analyzed using descriptive statistics with the pc -SAS program. The Linear Structural Relationship(LISREL) modeling process was used to find the best fit model which would predict the causal relationships among the variables. The overall fit of the hypothetical model to the data was moderate ［X$^2$＝69.6(df＝17, p＝.000), GFI ＝0.95, AGFI＝0.90, RMR＝0.087, NNFI＝0.86, NFI＝0.90］. The predictable variables, especially menstrual symptoms, physical symptoms and family function, had a significant direct effect on depression. but school life adaptation did not have a significant direct effect. These variables explained 18.1% of the total variance.