Lysimeter and field experiments were conducted in Sandy Loam to establish a simple estimation model for evapotranspiration (ET) of soybean for three years (l984-1986). Potential ET (PET) could be estimated by the eq.1 using Pan-evaporation (Eo) and was ranged from 1.1 to 4.6 mm/day during the experiments. PET (mm/day)=1.348＋0.573 Eo …(1) Crop coefficient (Kc=maximum ET/PET) could be estimated by the eq.2 using Growth degree (G=days after planting/total growing days) and was ranged from 0.2 to 1.1 and from 0.6 to 1.4 for monoculture cropping and double cropping followed by barley, respectively, during the experiments. Monoculture : Kc=0.016＋3.719 G-3.224 G2 …(2), Double cropping : Kc=0.609＋2.014 G-2.120 G2 …(2). However, the maximum Kc was shown when G was about 50% and 40% for the monoculture and the double cropping, respectively. Soil water coefficient (f=AET/maximum ET) could be estimated by the eq.3 using soil water tension (Ψ) in 15cm depth. and it was decleased to 0.2 when Ψ was 10 bar. f=0.755-0.537 log │Ψ│…(3) Consequentially, the model to estimate the Actual ET (AET) of soybean was determined as eq.4 with the correction coefficient of -0.380. AET(mm/day)=PETㆍKcㆍf -0.380 …(4) The estimated AET were compared with the measured AET to verify the model established above. The average deviation of the estimated ET(AET) was 0.5782~pm 0.338 (mm/day), and it would be within reasonable confidence range.