Six hundred and thirty smokers who intended to quit smoking themselves completed pre-cessation measures of self-efficacy, partner support, daily stresses and demographics. Subjects were contacted at 2, 7, 14, 30, 90 and 180 days post-cessation to determine smoking status and to re-administer the measures at 7, 14 and 30 days post-cessation. A series of logistic regression examined which prospective factors best predicted relapse between 0-2 days, 3-7 days, 8-14 days, 15-30 days, 31-90 days and 91-180 days. Relapse was predicted by different variables at different times; however, self-efficacy was a consistent predictor of relapse over time.