Augmentation of cue exposure (extinction) therapy with cognitive-enhancing pharmacotherapy may offer an effective strategy to combat cocaine relapse. To investigate this possibility at the preclinical level, rats and squirrel monkeys were trained to self-administer cocaine paired with a brief visual cue. Lever pressing was subsequently extinguished by withholding cocaine injections while maintaining response-contingent presentations of the cue. The glycine partial agonist D-cycloserine (DCS; 15 and 30 mg/kg in rats, 3 and 10 mg/kg in monkeys) was evaluated for its effects on the rate of extinction and subsequent reacquisition of cocaine self-administration. Compared with vehicle, pretreatment with 30 mg/kg DCS 0.5 h before extinction training reduced the number of responses and latency to reach the extinction criterion in rats, but neither dose of DCS altered these measures in monkeys. In both species, pretreatment with the higher dose of DCS before extinction training significantly attenuated reacquisition of cocaine self-administration compared with either extinction training in the absence of DCS or DCS in the absence of explicit extinction. Furthermore, treatment with 30 mg/kg DCS accompanied by brief handling (a stress induction) immediately after but not 6 h after extinction training attenuated reacquisition of cocaine self-administration in rats. No adverse effects of 10 mg/kg DCS were evident in quantitative observational studies in monkeys. The results suggest that DCS augmented consolidation of extinction learning to deter reacquisition of cocaine self-administration in rats and monkeys. The results suggest that DCS combined with exposure therapy may constitute a rational strategy for the clinical management of cocaine relapse.