Abstract: Parkinson’s disease (PD) is a neurological disorder characterized by rigidity, tremor, bradykinesia, and postural instability. Gait disturbance is one of cardinal symptoms of PD and affects the activities of daily living and quality of life. This symptom in advanced PD patients is usually refractory to medication and surgical intervention such as deep brain stimulation (Morishita et al., 2016). Therefore, physical therapy with an efficient exercise program is important to maintain or improve gait ability. Among the various rehabilitation programs, biofeedback therapy is drawing attention as a rehabilitation method. Biofeedback is a technique that aims to make unconscious or involuntary bodily processes perceptible so that patients can control them consciously. We recently reported the effect of biofeedback therapy using a robot suit hybrid assistive limb (HAL) on PD patients (Kotani et al., 2020). HAL is a robotic exoskeleton designed to facilitate movements and was developed based on the “interactive biofeedback” theory (Morishita and Inoue, 2016). Specifically, the movement of the robot is triggered by bioelectric signals detected by surface electrodes. It supports the spontaneous movement of impaired muscles and generates sensory feedback. In our study, we used the HAL lumbar type. Eight patients with advanced PD participated in this study. The participants performed HAL-assisted core exercises and squats, one session per day, for a total of five sessions. In our study, exercise with HAL improved walking ability after the short period of the five sessions, and the effect was maintained for 3 months. We considered that the HAL exercises were successful because the biofeedback of HAL is predominantly via the proprioceptive receptors. The robot enables patients to achieve repetitive movements.