The dynorphin/kappa opioid receptor (KOR) system has been implicated as one potential neurobiological modulator of the abuse-related effects of cocaine and as a potential target for medications development. This study determined effects of the KOR antagonist nor-binaltorphimine (nor-BNI) on cocaine self-administration under a novel procedure that featured two daily components: (1) a 2-hour ‘choice’ component (9:00–11:00 am) when monkeys could choose between food pellets and cocaine injections (0–0.1 mg/kg per injection, intravenous) and (2) a 20-hour ‘extended-access’ component (noon to 8:00 am) when cocaine (0.1 mg/kg per injection) was available under a fixed-ratio schedule to promote high daily cocaine intakes. Rhesus monkeys (n = 4) were given 14 days of exposure to the choice + extended-access procedure then treated with nor-BNI (3.2 or 10.0 mg/kg, intramuscular), and cocaine choice and extended-access cocaine intake were evaluated for an additional 14 days. Consistent with previous studies, cocaine maintained both a dose-dependent increase in cocaine choice during choice components and a high level of cocaine intake during extended-access components. Neither 3.2 nor 10 mg/kg nor-BNI significantly altered cocaine choice or extended-access cocaine intake. In two additional monkeys, nor-BNI also had no effect on cocaine choice or extended-access cocaine intake when it was administered at the beginning of exposure to the extended-access components. Overall, these results do not support a major role for the dynorphin/KOR system in modulating cocaine self-administration under these conditions in non-human primates nor do they support the clinical utility of KOR antagonists as a pharmacotherapeutic strategy for cocaine addiction.