The seasonal changes in evapotranspiration over a hill evergreen forest in northern Thailand (18¿48'N, 98¿54'E), in the Asian monsoon region, were simulated using a multilayer model and the boundary conditions above the canopy. The simulation considered the uncertainty in the leaf area index (LAI) and physiological parameters for both photosynthesis and stomatal conductance models. The parameters were based on the estimated LAI and determined by referring to the measured net photosynthesis rate and stomatal conductance for a single leaf. The simulated cumulative evapotranspiration and rainfall interception agreed with the values obtained from the water budget within these uncertainties. The sensitivity of these limits to both evaporation and transpiration was also investigated. The simulated transpiration peaked in the late dry season. The latent heat flux obtained with the eddy correlation technique showed that the forest continued to transpire in the late dry season. The heat pulse velocities also showed a peak in water use by individual trees in the late dry season. These results counter the view that evapotranspiration declines in the dry season, as has been reported previously for an evergreen forest and other vegetation in Thailand. The transpiration peak was thought to depend on the reduced wetness of the canopy, and the consequent lack of evaporation from it, and on the fact that there was little decline in stomatal conductance, even in the driest conditions.