Use of marijuana (Cannabis sativa) often begins in adolescence, and heavy adolescent marijuana use is often associated with impaired cognitive function in adulthood. However, clinical reports of long-lasting cognitive deficits, particularly in subjects who discontinue use in adulthood, are mixed. Moreover, dissociating innate differences in cognitive function from cannabis-induced deficits is challenging. Therefore, the current study sought to develop a rodent model of adolescent cannabinoid self-administration (SA), using the synthetic cannabinoid receptor agonist WIN55,212-2 (WIN), in order to assess measures of relapse/reinstatement of drug seeking and long-term effects on cognitive function assessed in a delay-match-to-sample working memory task and a spatial recognition task. Adolescent male rats readily self-administered WIN in 2-h or 6-h sessions/day, but did not demonstrate an escalation of intake with 6-h access. Rats exhibited significant cue-induced reinstatement of WIN seeking that increased with 21 days of abstinence (ie, 'incubation of craving'). Cognitive testing occurred in adulthood under drug-free conditions. Both 2-h and 6-h adolescent WIN SA groups exhibited significantly better working memory performance in adulthood relative to sucrose SA controls, and performance was associated with altered expression of proteins regulating GABAergic and glutamatergic signaling in the prefrontal cortex. Self-administered WIN did not produce either acute or chronic effects on short-term memory, but experimenter administration of WIN in adolescence, at doses previously reported in the literature, produced acute deficits in short-term memory that recovered with abstinence. Thus, SA of a rewarding cannabinoid in adolescence does not produce long-term cognitive dysfunction.