Individuals with mood/anxiety disorders may use cannabis for "self-medication," i.e., to induce positive mood or attenuate aversive mood states. However, little neurobiological evidence supports such use. The goal of this investigation was to test the hypothesis that cannabis use attenuates striatal response to reward in those with mood/anxiety disorders. Reward-related processing was measured using a monetary incentive delay task under functional MRI. Individuals with any lifetime mood/anxiety disorder diagnoses and problematic cannabis use ("Mood/Anxiety+CB"; n = 41) were compared with a propensity score-matched group of similar subjects without cannabis use ("Mood/Anxiety-CB"; n = 41), and a cannabis-naïve healthy control group (n = 35). Activations during win- and loss-anticipations were extracted from bilateral nucleus accumbens, dorsal caudate, and dorsolateral putamen. Mixed models were estimated for each region separately for win- and loss-anticipations, with a test for the main effect of group, condition (e.g., high-win, low-win, neutral), and their interaction. A significant main effect of group for win- and loss-anticipation was observed for each striatal region. Specifically, the Mood/Anxiety+CB group exhibited the lowest striatal activations across condition levels relative to both the Mood/Anxiety-CB and healthy group. A significant group-by-condition interaction was only observed for the dorsolateral putamen and indicated divergent activation modulation as a function of win and loss-magnitude for Mood/Anxiety+CB subjects. Finally, individuals with heavier recent cannabis use showed greater attenuation of gain-related activation in all three striatal regions. There was no such relationship for other illicit drugs. These data support the hypothesis that cannabis use in individuals with mood/anxiety disorders is associated with attenuated brain processing of reward magnitude, which may contribute to persistent affective symptoms.