The purpose of the present study was to determine whether or not the exercise intensity of water-walking for elderly women could be accurately prescribed by heart rate data obtained during treadmill exercise on land. Six healthy female volunteers, with a mean age of 62.2 ± 4.2 years, took part in this study. Walking on land was performed on a treadmill. Each subject completed three consecutive 4-minute walks at a progressively increasing velocity (40, 60 and 80 m·min^-1), with a 1-minute rest after both the first and second walks. The room temperature and relative humidity were 24.5 ± 0.2°C and 54.8 ± 4.0%, respectively. Walking in water was performed in a Flowmill, which is a treadmill positioned at the base of a water flume. Each subject completed three consecutive 4-minute walks at a progressively increasing belt and water-flow velocity (20, 30 and 40 m·min^-1), with a 1-minute rest after both the first and second walks. The water depth was at the level of the xiphoid process of each subject. The water temperature was 30.7 ± 0.1°C. The exercise intensity at the highest workrate was equivalent to 44.2 ± 10.3% of the heart rate reserve (HRR) during water-walking and 38.4 ± 4.7% of the HRR during land-walking. There was a highly significant linear relationship between heart rate (HR) and oxygen uptake (VO₂) during both water-walking and land-walking. The relationship between HR and VO₂ in both exercise modes was similar. Thus, the relationship of HR to VO₂ derived from a treadmill-graded walking test on land may be used to prescribe exercise intensity for water-walking in thermoneutral water.