Naltrexone reversibly blocks the effects of opioids and has been shown to decrease placebo analgesia. However, it is not clear (1) to what extent naltrexone affects pain modulation in a nontreatment context, for example, in response to pain cues or (2) how naltrexone given prior to pain-cue learning shapes pain responses. In a double-blind procedure prior to pain-cue conditioning, 30 healthy participants were randomized to receive an oral dose of naltrexone (50 mg) or inert pill. During functional magnetic resonance imaging, high and low pain pressures were paired with two different visual cues: a high pain cue and a low pain cue (learning sequence). During a test sequence, medium levels of pressure were used for both cues and the difference in subjective pain ratings following high and low pain cues was calculated. Results showed significant conditioned pain responses across groups (<.001); however, no significant difference between participants receiving naltrexone or inert pill (=.193). There was a significant correlation between the difference in high and low pain ratings during the learning sequence and the effect of high and low pain cues during the test sequence (r = .575, =.002). Functional magnetic resonance imaging analyses revealed no significant difference in brain activation between groups. Here, we demonstrate comparable learning of pain responses in participants treated with naltrexone or inert pill. The results point to the possibility that associative learning, and conditional responding to pain cues, is not dependent on endogenous opioids. Our results, using pain-cue conditioning to create reduced pain responses, contrast previous studies where opioid antagonists significantly reduced the placebo effect in treatment of pain.