Background: Identifying neural substrates that are differentially affected by drugs of abuse and natural rewards is key to finding a target for an efficacious treatment for substance abuse. Melanin-concentrating hormone is a polypeptide with an inhibitory effect on the mesolimbic dopamine system. Here we test the hypothesis that melanin-concentrating hormone in the lateral hypothalamus and nucleus accumbens shell is differentially involved in the regulation of morphine and food-rewarded behaviors. Methods: Male Sprague-Dawley rats were trained with morphine (5.0 mg/kg, subcutaneously) or food pellets (standard chow, 10-14 g) to induce a conditioned place preference, immediately followed by extinction training. Melanin-concentrating hormone (1.0 mu g/side) or saline was infused into the nucleus accumbens shell or lateral hypothalamus before the reinstatement primed by morphine or food, and locomotor activity was simultaneously monitored. As the comparison, melanin-concentrating hormone was also microinjected into the nucleus accumbens shell or lateral hypothalamus before the expression of food or morphine-induced conditioned place preference. Results: Microinfusion of melanin-concentrating hormone into the nucleus accumbens shell (but not into the lateral hypothalamus) prevented the reinstatement of morphine conditioned place preference but had no effect on the reinstatement of food conditioned place preference. In contrast, microinfusion of melanin-concentrating hormone into the lateral hypothalamus (but not in the nucleus accumbens shell) inhibited the reinstatement of food conditioned place preference but had no effect on the reinstatement of morphine conditioned place preference. Conclusions: These results suggest a clear double dissociation of melanin-concentrating hormone in morphine/food rewarding behaviors and melanin-concentrating hormone in the nucleus accumbens shell. Melanin-concentrating hormone could be a potential target for therapeutic intervention for morphine abuse without affecting natural rewards.