Objective: To identify the differences in food habits and preferences among the adolescents according to socio-demographic characteristics. Design: Cross-sectional, cluster design survey in 2002. Setting: Eight middle schools in two distinct socio-economic areas of the Jiangsu province, China. Subjects: Some 824 young adolescents (12-14 y) attending public schools with a response rate of 99%. Methods: A self-administered questionnaire containing questions on food and meal frequencies, food preferences and sociodemographic characteristics was used. Results: High socio-economic status (SES) and urban residence was positively associated with intake of high-energy foods, such as foods of animal origin, Western style foods and dairy products. In all, 76% of the students had three meals a day regularly , but 8.1% urban students vs 3.4% rural students had breakfast only 1-3 times per week or less often. Daily fruit consumption was fairly common, but with clear differences by SES. Only about 42% of the boys and 55% of the girls from low SES families ate fruit daily, compared with 66% and 72%, respectively in the high SES families. Urban boys had the lowest proportion of daily consumers of vegetables (67.0%). More urban students drank milk daily than the rural students (68.7 vs 38.5%). The frequency of milk drinking also showed a strong positive association with SES. About 10% of the high SES boys consumed hamburgers daily compared with 2.8% of the low SES boys. More than half of the students reported a liking for Western style fast foods including hamburgers, soft drinks and chocolate. Among high SES boys, 21.5% consumed soft drinks on a daily basis; however, as many as 72.3% wanted to drink soft drinks more often if they could afford it. Conclusions: SES and urban location were positively associated with frequency of intake of high-energy foods. Reported food preferences may enforce this trend. Nutrition education for adolescents and parents is needed to promote healthy eating. Health Authorities should strengthen the monitoring of food intake and its association with overweight/obesity.