Data from the Working Well trial (n = 2379) were used to test the capacity of 19 variables to predict smoking cessation at 1- and 2-year follow-ups. Among the core constructs of transtheoretical model (TTM), stage of change was the best predictor. The processes of change and the pros and cons of smoking were relatively ineffective predictors. Among other variables, self-efficacy, cigarettes per day, duration of longest quit attempt during the previous year, and the contemplation ladder were the most effective stand-alone predictors. A composite of cigarettes per day and quit duration was particularly effective for predicting cessation. Consistent with Farkas et al. (Farkas AJ, Pierce JP, Zhu SH, Rosbrook B, Gilpin EA, Berry C, Kaplan RM, Addiction 91:1271–1280, 1996), multivariate analyses including the composite variable (cigarettes per day and quit duration) and the stages of change revealed the composite variable to be the better predictor of cessation.