To study the effectiveness of a very brief advice (30 s) on smoking cessation. A proof-of-principle single-blind, randomized controlled trial (RCT). Medical outpatient clinics of a general hospital in Guangzhou, China. One hundred and twenty-six male current smokers randomly allocated into an intervention (n 74) and a control group (n 52). A health warning by physicians that half of all smokers would be killed by smoking, an advice to quit immediately and referral to a cessation clinic. The control group received none. Primary: seven-day quitting point prevalence at 6 months. Secondary: 7-day point prevalence at 1, 3 and 12 months, sustained abstinence at 3, 6 and 12 months, smoking reduction by half and cessation clinic attendance. By intention-to-treat analysis, 7-day quitting point prevalence rates at four follow-ups were 27.0, 23.0, 21.6 and 18.9 in the intervention group, compared with 5.8, 3.8, 5.8 and 5.8 in the control group (first three P 0.05). At 3, 6 and 12 months, sustained abstinence prevalence rates were 18.9, 17.6 and 14.9 versus 3.8, 3.8 and 3.8 (P 0.035, 0.046, 0.074). More smokers in the intervention group had reduced smoking. Almost no participants attended the cessation clinic. Our findings support the need for large RCTs on minimal interventions with the one in two warning.