The relationship between pain severity and the extent of injury to a peripheral nerve remains elusive. In this study, we compared the pain behavior resulting from partial (1/3–1/2 thickness) and full L5 spinal nerve ligation (SNL) in rats. The decrease in paw withdrawal threshold (PWT) to mechanical stimuli in the hindpaw ipsilateral to the injury was comparable in the two groups on days 3–21 post-injury. However, the decreased PWT recovered earlier in the partial SNL group than in the full SNL group. These observations suggest that the duration of neuropathic pain behavior, but not the early development of mechanical allodynia, is dependent on the extent of nerve injury. On days 6 and 15 post-injury, when the mechanical allodynia was similar in the two groups, systemic morphine induced a greater reduction of mechanical allodynia in the partial SNL group than in the full SNL group. Furthermore, in partial SNL rats, at post-injury time points when they had largely recovered from the neuropathic pain state, systemic administration of naloxone hydrochloride (day 53) or naloxone methiodide (a non-selective peripherally acting opioid receptor antagonist; day 64) or intra-plantar injection of naloxone methiodide rekindled mechanical pain hypersensitivity in the ipsilateral hindpaw, suggesting a prolonged activation of endogenous opioidergic pain-inhibition. Therefore, partial SNL in rats may represent an efficient model for studying the mechanisms of neuropathic pain, testing effects of analgesic/antihyperalgesic drugs, and understanding endogenous pain-inhibitory mechanisms that lead to reversal of the pain behavior with time.