One hundred one smokers were divided into high and low trait anxiety groups on the basis of a normalized score on the Profile of Mood States Anxiety/Tension Scale and were randomly assigned to receive buspirone or placebo in a double-blind fashion. After a 1-week baseline, smokers were exposed to an 8-week drug and behavioral intervention involving buspirone or placebo (up to 60 mg/day) with concurrent group cognitive behavioral intervention. All smokers were to quit smoking on the target date, set at 4 weeks after the program began. Medication was provided for an additional 4 weeks after group treatment ended. The results showed that buspirone had a beneficial effect on smoking abstinence but only among smokers who were already relatively high in anxiety and only for as long as the drug was available. Moreover, when provided to smokers who were relatively low in anxiety, the drug appeared to interfere with abstinence, although these effects also reversed when the drug was withdrawn. These effects were associated with an attenuation of the expected rise in anxiety before the quit date and its actual reversal thereafter, but only in the buspirone high-anxiety group. One-month abstinence averaged 88, 61, 60, and 89% for the buspirone high-anxiety, placebo high-anxiety, buspirone low-anxiety, and placebo low-anxiety groups, respectively. By 12 months, abstinence for the buspirone and placebo high- and low-anxiety groups fell to 12, 23, 41, and 36%, respectively. No differences were observed for measures of self-efficacy, symptoms of withdrawal, medication side effects, or compliance.