Nicotine replacement therapy (NRT) repeatedly has been shown to improve smoking treatment outcome. The major mechanism posited for this improvement in outcome is that NRT reduces nicotine craving and withdrawal. The authors tested this hypothesized mechanism of action using real-time data on craving and withdrawal, collected by ecological momentary assessments administered on a palm-top computer. Smokers (N = 324) were randomized to receive either active high-dose (35 mg) 24-hr patches or placebo. Increases in positive affect and decreases in craving, negative affect, and attention disturbance severity were related to lower risk of lapsing. Although NRT treatment did significantly decrease withdrawal and craving severity, these reductions only partially accounted for NRT's impact on time to first lapse: The results from a mediation analysis showed that the hazard ratio for NRT, when controlling for withdrawal and craving severity, was only a third to a half lower than the uncontrolled hazard ratio for NRT alone. This suggests that other mechanisms for the effectiveness of NRT need to be examined.