The effectiveness of worksite interventions to reduce smoking needs to be enhanced because randomized controlled trials to date have produced mixed findings. The present study tested the ability of social-cognitive variables to mediate the past behavior-future behavior relationship and the effectiveness of implementation intentions to break the past behavior-future behavior relationship in a brief theory-based worksite intervention designed to reduce smoking. Smoking behavior and psychosocial orientation to quit (operationalized by theory of planned behavior variables and temptations) were measured at baseline; participants (N = 90) randomized to the experimental condition were also asked to form an implementation intention in their place of work. Identical measures taken 2 months postbaseline revealed that intention was a potent mediator of the past behavior-future behavior relationship. More important, significantly more people quit smoking in the experimental condition than in the control condition. Decomposition of these effects showed that implementation intentions worked best for individuals who were more motivated to quit at baseline and suggest that harnessing both motivational and volitional processes might enhance the effectiveness of worksite smoking cessation programs.