Introduction The hybrid assistive neuromuscular dynamic-stimulation (HANDS) therapy is closed-loop electromyography-controlled neuromuscular electrical stimulation (NMES) with a wrist-hand splint for patients with moderate to severe hemiparesis even including chronic stroke. It is repeatedly reported HANDS therapy induce significant recovery in upper extremity motor function for relatively severe paresis with a minimal ability for volitional muscle activation, who do not meet the criteria for other intensive therapies such as constrained-induced movement therapy. Although almost the half of stroke survivors are suffering from somatosensory impairment, and it is indicated that the sensory recovery would have a certain beneficial impact on motor functional recovery, neither the detailed mechanisms of somatosensory recovery nor the interaction with motor recovery are not determined enough especially for chronic stroke. On these ground, we have investigated the effect of HANDS therapy on the feature of sensory recovery. Methods For 23 chronic stroke patients with hemiplegia, 3 weeks inpatient HANDS therapy intervention was taken place. Median and Tibial nerve Sensory Evoked Potential (SEP) and behavioral sensory assessments including Semmes-Weinstein monofilament test and thumb-localizing test were performed at pre and posttreatment, as well as the motor functional assessments. Results A significant increase in the number of SEP peaks in the Median nerve and a significant reduction in the latency gap between N20 and P45 were observed in the paretic side after the intervention, whereas no significant difference was observed in any of these behavioral assessments. Although motor functional recovery was observed as in the previous studies, there was no significant correlation detected between motor function and each sensory related indices including SEP peak number and behavioral assessments. Conclusion Our result firstly suggests that closed-loop electromyography-controlled NMES training induce a sort of plastic changes especially in the level of cortex even in the chronic stroke patients. SEP would be a good evaluation modality which has higher sensitivity to detect a smaller plastic change induced by rehabilitative intervention.