The opioid system is implicated in the hedonic and motivational processing of food, and in binge eating, a behaviour strongly linked to obesity. The aim of this study was to evaluate the effects of 4 weeks of treatment with the mu-opioid receptor antagonist GSK1521498 on eating behaviour in binge-eating obese subjects. Adults with body mass index ⩾30 kg m−2 and binge eating scale scores ⩾19 received 1-week single-blind placebo run-in, and were then randomized to 28 days with either 2 mg day−1 GSK1521498, 5 mg day−1 GSK1521498 or placebo (N=21 per arm) in a double-blind parallel group design. The outcome measures were body weight, fat mass, hedonic and consummatory eating behaviour during inpatient food challenges, safety and pharmacokinetics. The primary analysis was the comparison of change scores in the higher-dose treatment group versus placebo using analysis of covariance at each relevant time point. GSK1521498 (2 mg and 5 mg) was not different from placebo in its effects on weight, fat mass and binge eating scores. However, compared with placebo, GSK1521498 5 mg day−1 caused a significant reduction in hedonic responses to sweetened dairy products and reduced calorific intake, particularly of high-fat foods during ad libitum buffet meals, with some of these effects correlating with systemic exposure of GSK1521498. There were no significant effects of GSK1521498 2 mg day−1 on eating behaviour, indicating dose dependency of pharmacodynamics. GSK1521498 was generally well tolerated and no previously unidentified safety signals were detected. The potential for these findings to translate into clinically significant effects in the context of binge eating and weight regain prevention requires further investigation.