Peripheral nerve disorder of the lower extremities causes drop foot and disturbs the daily living activities of patients. The ankle joint hybrid assistive limb (HAL) provides voluntary ankle joint training using surface bioelectrical signals from the muscles of the lower extremities. We investigated the neurological effects of ankle joint HAL training in three patients. Sensory nerve action potentials (SNAPs) and compound muscle action potentials (CMAPs) were analyzed for the peroneal and tibial nerves prior to the first ankle joint HAL training session. Integrated surface electromyography EMG signals were recorded before and after the HAL training sessions to evaluate the effects of training for neuromuscular disorders. The patients were hospitalized to receive rehabilitation with HAL training for 2 weeks. The HAL training was performed daily with two 60 min sessions. All cases demonstrated severe neuromuscular impairment according to the result of the CMAP. All integrated EMG measurements of antagonistic muscle activities decreased after the ankle joint HAL training. The manual muscle testing (MMT) scores of each muscle were slightly increased after the HAL intervention for Case 2(tibialis anterior, from 2 to 2+; gastrocnemius muscles, from 2- to 2; extensor digitorum longus, and extensor hallucis longus, from 1 to 3). The MMT scores were also slightly increased except for gastrocnemius muscle for Case 3 (tibialis anterior, extensor digitorum longus, and extensor hallucis longus, from 2- to 2). These two patients demonstrated voluntary muscle contractions and nerve signals in the CMAP before the HAL training. Even though the amplitude of CMAPs was low, the HAL training may provide voluntary ankle joint movements by reducing the antagonistic muscle contraction via computer processing. The HAL training may enhance muscle movement and coordination through motor learning feedback.