Smokers' reactions to a laboratory assessment that simulated high-risk-for-smoking-relapse situations were monitored prior to and at the end of treatment for smoking cessation. Measures included self-reported urges to smoke, efficacy, anxiety, behaviorally rated coping effectiveness, and heart rate. Observed pretreatment responses were unrelated to smoking outcome at the end of treatment. Univariate analyses indicated that decreased efficacy and coping effectiveness and increased urges at the end of treatment were related to relapse during 6-month follow-up. However, multivariate analysis demonstrated that these relationships were a function of the end-of-treatment smoking level. When end-of-treatment smoking rate was controlled for statistically, only increased heart rate response during the assessment predicted smoking status at 6 months. The theoretical significance of the findings is discussed.