Previous studies suggest that GABA-B receptor agonism may represent an effective pharmacological approach to treat addictive disorders. Baclofen is a selective GABA-B receptor agonist which has been investigated as a potential treatment for alcohol use disorder. However, research is needed to understand the biobehavioral mechanisms underlying baclofen's effect on alcohol use. In the present randomized, double-blind, placebo-controlled study, thirty-four alcohol-dependent individuals were randomized to receive baclofen (30 mg/d) or placebo for a week, and then participated in a laboratory experiment consisting of three procedures: alcohol cue-reactivity, priming, and self-administration. During the experiment, craving and other subjective responses to alcohol were assessed, and blood samples were collected for pharmacokinetic measurements. The effects of baclofen on the relationships between different alcohol-related laboratory parameters were investigated. Baclofen pharmacokinetic parameters and their correlations with behavioral measures were also examined. Results showed that baclofen disrupted the link between alcohol priming and self-administration, as indicated by significant interaction effects between drug condition (baclofen vs. placebo) and some of the priming variables (alcohol craving: F3,9 = 6.03, p = 0.01; alcohol sedation: F3,6 = 7.16, p = 0.01) on the total amount of alcohol self-administered. Considerable interindividual variability in baclofen pharmacokinetic parameters was observed. Maximum plasma concentrations of baclofen negatively correlated with cue-induced alcohol craving (r = -0.57, p = 0.03) and priming-induced ratings of 'like more' (r = -0.59, p = 0.02). In conclusion, baclofen may work by dissociating the link between an initial drink (priming) and subsequent alcohol consumption (self-administration). Considerable pharmacokinetic variability is an important factor to take into account when employing baclofen as a treatment for alcohol use disorder.