This study aimed to explore the relationship between the family structures in which adolescents lived and their health-related behavior. This study, applying a cross-sectional descriptive design, investigated the hypothesis that family structure and level of parent education are strongly associated with health-related behavior in adolescents. Study subjects consisted of sixth to eighth graders drawn from six schools in Taipei and Kaohsiung between September 2002 and March 2003. Five hundred and forty-six youths, with a mean age of 13 years, participated in this study. Data were analyzed using descriptive statistics, Student's t test, and one-way ANOVA. Multiple stepwise regression was used to determine relationships between health promotion scores and family structure, parent level of education, gender and other variables. Results indicated that, when potentially confounding factors were controlled, adolescents living with both parents have healthier behavior than those living with a single parent. In addition, when potentially confounding factors were controlled, adolescents residing with both parents with relatively high educational levels only achieved higher healthy behavior scores in the dimensions of health responsibility, nutrition, and exercise. Further, adolescent females had substantially higher health-related behavior scores than their male peers in all dimensions, with the exception of "exercise behavior". Findings support the hypothesis that family structure and female gender are significant variables affecting adolescent health-related behavior. Research results are useful in developing intervention strategies to assist adolescents in single parent families.