The effectiveness of a relapse prevention program was studied in a population of 744 smokers. Subjects first attended an intensive 4-day series of 2-hr meetings in which they were trained to use more than 40 behavioral and cognitive smoking cessation techniques. At the 1-week follow-up session, those abstinent from smoking (79%, carbon monoxide verified) were randomly assigned to one of three follow-up conditions: (a) a three-session skills-training program in which subjects developed and actively rehearsed individually tailored coping strategies for likely relapse situations, (b) a three-session discussion control condition in which subjects discussed maintenance but did not develop or rehearse coping strategies, or (c) a no-treatment control condition. Survival analysis indicated higher abstinence rates for the skills-training group than for the control groups throughout the following year. After 12 months, the biochemically confirmed continuous abstinence rate was higher in the skills-training group (41.3%) than in the discussion and no-treatment groups (34.1% and 33.3%, respectively).