Our aim was to evaluate the effects of a 4-week training program using a self-powered pedaling wheelchair on brain perfusion in patients presenting with lower limb hemiparesis due to stroke, brain injury, or spinal cord injury. Our cross-sectional observational study included seven patients with lower limb hemiparesis (five men, two women; mean age, 68.3 ± 17.5 years), due to the following causes: cerebral hemorrhage (n = 1), stroke (n = 4), brain contusion (n = 1), and spinal cord injury (n = 1). The control group consisted of eight healthy participants (3 men, 5 women; mean age 62 ± 8 years). The training program consisted of five bouts of 3-min continuous pedaling per day (total, 15 min/day). The outcome variable of interest was blood flow velocity in the middle cerebral artery (time average peak [TAP], cm/s) measured using Doppler. TAP was measured at rest and after a 3-min pedaling bout, before and after the training program. In the patient group, TAP was significantly greater after the 3-min bout than at rest, both before and after the training program (p < 0.05). There was no effect of pedaling identified in the control group. In the patient group, TAP increased significantly (p < 0.05) after training, both at rest (36.9 ± 16.9 to 47.6 ± 13.8 cm/s), and after the 3-min bout (43.3 ± 13.3 to 50.5 ± 15.1 cm/s). Our pedaling wheelchair provided a safe and effective intervention to improve brain perfusion in this patient population.