BACKGROUND: Anti-smoking advice from general practitioners has proven efficacy. However, general practitioners do not exploit a large proportion of opportunities to discuss smoking with patients. AIM: A study aimed to explore general practitioners attitudes towards discussing smoking with patients and to assess how these influence the quantity of anti-smoking advice that general practitioners report giving during routine consultations. It also aimed to determine the extent to which general practitioners report using evidence-based interventions against smoking and to discover the problems they experience when discussing smoking with patients. METHOD: A postal survey of all 468 general practitioners on the Leicestershire Family Health Services Authority list was conducted. General practitioners' attitudes were assessed by scoring 13 attitude statements using a six-point Likert-type scale. They were also asked to rank (from a list of 12 items) the five approaches that they found most productive and (from a list of 11 items) the five problems that they most commonly encountered when giving anti-smoking advice to patients. RESULTS: A total of 327 questionnaires (70%) were returned. Most respondents (97%) thought that their advice was more effective when linked to patients' presenting problems and 65% reported that linking their anti-smoking advice to patients' presenting complaints was one of their three most preferred approaches to discussing smoking. Advising all presenting smokers to quit was considered by 40% of respondents to be an appropriate use of time but 76% reported that patients' lack of motivation was one of the three most commonly encountered problems. An analysis of the ratings of the 13 statements suggested that general practitioners who reported the greatest smoking cessation activity during routine consultations held more positive attitudes towards discussing smoking with patients. CONCLUSION: This study suggests that general practitioners believe that their anti-smoking advice is more effective when linked to patients' presenting complaints, and this belief appears to be reflected in the way in which general practitioners approach smoking cessation with patients. The findings may indicate that general practitioners are unlikely to accept a role in a population-based anti-smoking strategy which demands that they discuss smoking with all presenting smokers.