The purpose of this study was to assess the test-retest reliability of heart rate (HR) and velocity measurements during peak effort and free treadmill walking tests in older patients with gait-impaired chronic hemiparetic stroke and control group. Twenty-two adults (13 men, 9 women; mean age, $73.7{pm}5.2$ yrs) with chronic hemiparetic stroke are the experimental group. Nineteen elderly people (5 men, 14 women; mean age, $72.3{pm}3.5$ yrs) were recruited as control group. Patients had mild to moderate chronic hemiparetic gait deficits, making handrail support necessary during treadmill walking. Free and peak effort treadmill walking tests were measured and then repeated at least two days later. Reliability was calculated from HR and walking velocity during free and peak effort treadmill walking test. Among the people who had strokes, HR [ICC(2,1)=.85, r=.86] and velocity [ICC(2,1)=.93, r=.93] were good parameters during free testing. Maximal testing generated good results for HR [ICC(2,1)=.81, r=.82] and velocity [ICC(2,1)=.96, r=.96] with the chronic hemiparetic stroke. In elderly people, HR [ICC(2,1)=.59, r=.62] and velocity [ICC(2,1)=.77, r=.76] were moderately reliable during free testing. Maximal testing produced moderate parameters for HR [ICC(2,1)=.74, r=.74] and velocity [ICC(2,1)=.66, r=.66] in the elderly. This study provides that free and maximal treadmill testing produce highly reliable HR and velocity measurements in adults with chronic hemiplegia using minimal handrail support.