To clarify the determining processes of stable isotope ratios of hydrogen (δD) and oxygen (δ18O) in paddy water and the factors affecting them, we collected irrigated water, drain water, and ponded water in mid-May, early in June, and at the beginning of August in 2002 from a paddy field, where intensive measurement of water balance was performed. The values of δD and δ18O gradually increased along the flow direction of paddy water. In a δ18O vs. δD diagram, these data were plotted linearly for each month, and the slopes ranged from 4.3 to 5.8, showing kinetic isotope fractionation during evaporation. In August, the increase of δ values was much less than that in May and June, while the slope of evaporation line (SEL) was largest. Considering evaporative fractionation based on the Craig-Gordon model, the large SEL in August resulted from limited kinetic fractionation due to high humidity under a grown rice canopy. The grown canopy would also lower the increase of δ values by restricting evaporation. At the beginning of August, however, rainfall from a thunderstorm and the change of irrigation management also influenced δ values, so the lowered δ-increase was not attributed only to the restriction of evaporation. The result of the study shows the possibility to separate evaporation and transpiration using isotope hydrological approaches, which is usually impossible by micrometeorological direct measurements.