Men who have sex with men (MSM) experience disproportionately high rates of HIV and substance use. Though interventions based on cognitive behavioural therapy (CBT) are effective for many psychiatric disorders, systematic reviews of CBT for substance use present mixed results and CBT has not been systematically reviewed for HIV risk reduction in MSM. Randomised controlled trials (RCTs) of non-pharmacological CBT for substance-using MSM were systematically reviewed to assess the evidence for these interventions' impact on unprotected anal intercourse (UAI). We searched 20 databases on 22 September 2012. Of the 2801 records identified, both authors separately assessed 70 full-text articles for inclusion. In three included RCTs, one RCT compared CBT against non-treatment control and two RCTs compared CBT against CBT variants. Across studies, CBT led to UAI reductions, though at one-year follow-up, these reductions were generally not greater than in other trial arms. This evidence of moderate quality indicates that while CBT may reduce UAI in substance-using MSM, it is unclear if CBT is more effective than mere assessment. All RCTs tested intensive, clinic-based interventions requiring sustained participant commitment, indicating a need for research into briefer interventions, with particular focus on episodic substance use and effectiveness trials.