The global prevalence of obesity has been steadily increasing. Although pharmacotherapy and bariatric surgeries can be useful adjuvants in the treatment of morbid obesity, they may lose long-term effectiveness. Obesity result largely from unbalanced energy homeostasis. Palatable and densely caloric foods may affect the brain overlapped circuits involved with homeostatic hypothalamus and hedonic feeding. Deep brain stimulation (DBS) consists of delivering electrical impulses to specific brain targets to modulate a disturbed neuronal network. In selected patients, DBS has been shown to be safe and effective for movement disorders. We review all the cases reports and series of patients treated with DBS for obesity using a PubMed search and will address the following obesity-related issues: (i) the hypothalamic regulation of homeostatic feeding; (ii) the reward mesolimbic circuit and hedonic feeding; (iii) basic concepts of DBS as well as the rationale for obesity treatment; (iv) perspectives and challenges in obesity DBS. The small number of cases provides preliminary evidence for the safety and the tolerability of a potential DBS approach. The ventromedial (
n
= 2) and lateral (
n
= 8) hypothalamic nuclei targets have shown mixed and disappointing outcomes. Although nucleus accumbens (
n
= 7) targets were more encouraging for the outcomes of body weight reduction and behavioral control for eating, there was one suicide reported after 27 months of follow-up. The authors did not attribute the suicide to DBS therapy. The identification of optimal brain targets, appropriate programming strategies and the development of novel technologies will be important as next steps to move DBS closer to a clinical application. The identification of electrical control signals may provide an opportunity for closed-loop adaptive DBS systems to address obesity. Metabolic and hormonal sensors such as glycemic levels, leptin, and ghrelin levels are candidate control signals for DBS. Focused excitation or alternatively inhibition of regions of the hypothalamus may provide better outcomes compared to non-selective DBS. Utilization of the NA delta oscillation or other physiological markers from one or multiple regions in obesity-related brain network is a promising approach. Experienced multidisciplinary team will be critical to improve the risk-benefit ratio for this approach.