Dietary pattern analysis has emerged as important instruments to identify modifiable dietary risk factors for non-communicable diseases. The aim of this study was to determine the major dietary patterns among adolescents in Petaling District, Selangor and their associations with socio-demographic characteristics and obesity. An analytic cross- sectional study design was conducted in selected secondary schools in Petaling District. Sampling with probability proportionate to size was used and five schools were selected. Self-administered semi-quantitative food frequency questionnaire was used for data collection. Weight was measured with a digital bathroom scale (TANITA model) and height was measured using SECA body meter. Principal component factor analysis using varimax orthogonal transformation was used to identify the dietary patterns. Chi square was used to test for associations of dietary patterns with socio-demographic characteristics and obesity. Three major dietary patterns were identified: firstly, fruits and vegetables; secondly, sugar and fatand finally, meat and chicken which explained for 12.7%, 11.6% and 10.7% variation in food intake, respectively. There were significant associations between ethnicity, religion, family income, educational level of parents and the dietary patterns. However, there was no significant association between obesity and the dietary patterns. It may be more effective to describe a healthy diet using results of dietary pattern analysis in public health intervention, rather than describing single food items or nutrients. It is recommended that nutrition education programmes should be implemented in schools so as to prevent the development of obesity in the non-obese.