ABSTRACT: Addiction is characterized by high relapse susceptibility, and relapse can be triggered by drug-associated cues. Cue presentation induces retrieval of the drug-cue memory, which becomes labile and must be reconsolidated into long-term storage. Repeated unpaired cue presentation, however, promotes extinction. Cue-reactivity can be reduced by blocking reconsolidation or facilitating extinction, which are mediated by NMDA receptors (NMDArs). However, the role of NMDArs in either process following self-administration is unclear. Thus, to determine their role in extinction, rats learned to self-administer cocaine before receiving injections of the NMDAr antagonist CPP immediately after four 45-min extinction sessions. During a subsequent 90-min extinction retention test, CPP-treated rats lever pressed less than saline-treated rats indicating that NMDAr blockade facilitated extinction or disrupted drug-cue memory reconsolidation. In addition, infusing CPP into the infralimbic medial prefrontal cortex (IL-mPFC), a structure implicated in extinction, before four 45-min or immediately after four 30min extinction sessions, had similar results during the extinction retention tests. Next, the GluN2A-selective antagonist NVP or GluN2B-selective antagonist Ro25 was infused into IL-mPFC or nucleus accumbens (NAc) shell, another structure implicated in extinction, after four 45-min extinction sessions. Blocking GluN2A-, but not GluN2B-, containing NMDArs, in IL-mPFC or NAc shell reduced lever pressing during the extinction retention tests. Finally, to dissociate reconsolidation from extinction, NVP was infused into IL-mPFC after four 10-min reactivation sessions, which resulted in reduced lever pressing during the retention test. These results indicate that IL-mPFC GluN2A-containing NMDArs modulate reconsolidation, and suggest a novel treatment strategy, as reducing cue reactivity could limit relapse susceptibility.