Invasive neurostimulation therapies may be proposed to patients with neuropathic pain refractory to conventional medical management, in order to improve pain relief, functional capacity, and quality of life. In this review, the respective mechanisms of action and efficacy of peripheral nerve stimulation (PNS), nerve root stimulation (NRS), spinal cord stimulation (SCS), deep brain stimulation (DBS), and motor cortex stimulation (MCS) are discussed. PNS appears to be useful in various refractory neuropathic pain indications (as long as there is some preservation of sensation in the painful area), such as intractable chronic headache, pelvic and perineal pain, and low back pain, but evidence for its efficacy is not strongly conclusive, and large-scale randomized controlled studies are necessary to confirm the efficacy in the long term. Spinal cord stimulation (SCS) has been validated for the treatment of selected types of chronic pain syndromes, such as Failed Back Surgery Syndrome, and Complex Regional Pain Syndrome type I. When neuropathic pain is secondary to a brain lesion (especially following stroke) or a trigeminal lesion, stimulation of brain structures is required. Deep brain stimulation (DBS), which can be proposed with targets like the periventricular/periaqueductal gray matter or the sensory thalamus, is increasingly replaced by motor cortex stimulation (MCS), mainly because it is safer, more easily performed, and probably more effective in a wider range of indications (including central post-stroke pain). The respective places of DBS and MCS in some selected indications, such as peripheral neuropathic pain and phantom limb pain, have yet to be clearly delineated. Controlled trials, with the stimulator switched ON or OFF in a double-blind procedure, have demonstrated the efficacy of MCS in the treatment of peripheral and central neuropathic pain, although these trials included a limited number of patients and need to be confirmed by large, controlled, multicenter studies. Despite technical progress in neurosurgical navigation, guided by neuroimaging and intraoperative electrophysiology to optimize electrode positioning, MCS results are still variable, and validated criteria for selecting good candidates for implantation are lacking, except clinical response to preoperative rTMS, which showed correlations with a good response to MCS-induced analgesia. However, the evidence in favor of this technique is sufficient to include it in the range of treatment options for refractory neuropathic pain.