Background: The aim was to investigate the relative effectiveness of two strategies outlined in national smoking cessation guidelines for the treatment of tobacco dependence in general practice: in-practice management vs. referral to specialist help. Method: Cluster randomised controlled trial in which GPs were randomised to one of two intervention arms; (1) standard GP management (in-practice), or (2) referral to a specialist triage system that allowed smokers to choose from two alternative forms of assistance provided by the Victorian Quitline (callback counselling or the Internet-based Quit Coach). Between March and December 2004, 45 primary care physicians from 44 practices in Victoria, Australia recruited 1,052 adult current smokers. The primary outcome measures were point prevalence abstinence at 3 and 12 months; sustained abstinence was also measured. Results: Using a conservative analysis (missing data coded as treatment failure), point prevalence abstinence was reported by significantly more participants in the referral group (15.2%) than the in-practice group (8.9%) at 12 month follow-up; OR =1.83 (95% CI, 1.0 – 3.2). Group differences for sustained abstinence at 12 months, were marginally significant (p = 0.07). Among those in the referral condition, point prevalence and sustained abstinence was associated with the amount of assistance received. Conclusions: More smokers quit in the referral condition than in the in-practice condition, because they received more help from outside the practice, while getting the same amount of help within it. Implications: This study provided evidence that GPs referring smokers to an evidence-based cessation service can result in increased cessation, and is acceptable to both GPs and smokers.