The present study investigated the longitudinal impact of the Project PATHS, a large-scale curriculum-based positive youth development program in Hong Kong, on the development of adolescents' risk behavior over a period of five years. Using a longitudinal randomized controlled design, eight waves of data were collected from 19 experimental schools in which students participated in the Project PATHS (N = 2,850 at Wave 8) and 24 control schools without joining the Project PATHS (N = 3,640 at Wave 8). At each wave, students responded to measures assessing their current risk behaviors, including delinquency, use of different types of drug, and their intentions of participating in risk behaviors in the future. Results demonstrated that adolescents receiving the program exhibited significantly slower increases in delinquent behaviors and substance use as compared to the control participants. During two years after the completion of the program, differences in youth risk behaviors in the two groups still existed. These results suggest that the Project PATHS has long-term effect in preventing adolescent problem behavior through promoting positive youth development.