Physicians used either an autonomy-supportive or a controlling interpersonal style to counsel smokers based on National Cancer Institute guidelines. Physician autonomy support was rated from audiotapes, and patients' perceived competence and autonomous motivation for quitting were self-reported on questionnaires. Validated point prevalences for 6, 12, and 30 months and for continuous cessation were examined. The intervention did not have a direct effect on quit rates; however, structural equation modeling supported the self-determination process model of smoking cessation. The model indicated that the autonomy-supportive intervention was rated as more autonomy supportive, that rated autonomy support predicted autonomous motivation, and that autonomous motivation predicted cessation at all points in time. Perceived competence contributed independent variance to cessation only at 6 months.