It has been postulated that opioid systems in the brain may play a role in ethanol reinforcement. In this respect, μ- and δ-opioid receptors may mediate the rewarding effects whereas κ receptors are thought to mediate the aversive effects of opioids. Accordingly, long-acting benzomorphans such as bremazocine, that simultaneously act as μ and δ receptor antagonists and κ receptor agonists may be particularly effective in reducing ethanol self- administration. Therefore, we studied the effect of bremazocine on oral ethanol self-administration in rats using a paradigm [unrestricted free- choice drinking of 10% (v/v) ethanol], previously shown to cause long-term neuroadaptations in the nucleus accumbens and caudate putamen. Bremazocine (0.1 mg/kg, once daily for five consecutive days) reduced ethanol drinking by about 50% during the active period of the animals, whereas the intake of sucrose (3-10% w/v) was affected neither in naive nor in ethanol-experienced rats. This effect of bremazocine appeared not to be secondary to its acute sedative effect or the slight increase in total fluid consumption. Unlike bremazocine, the selective κ-opioid receptor agonist U50,488H (10mg/kg, once daily) inhibited ethanol drinking only during the first of 5 treatment days and the opioid receptor antagonist naltrexone (0.3-10 mg/kg, once daily) only caused a modest (about 20%) suppression of ethanol drinking during the first hours after drug injection. Thus, bremazocine appears to be far more potent than the clinically applied drug naltrexone in this respect. Our data further support the role of opioid receptors in ethanol reinforcement and indicate that long-acting mixed-action opioids such as bremazocine may be useful as adjuvants for the clinical management of ethanol addiction.