Abstract: To assess the applicability of ecological model for simulating evapotranspiration and soil water content in farmland, BEPS (Boreal Ecosystem Productivity Simulator), a processbased ecological model, was employed to simulate daily evapotranspiration (ET) and soil water dynamics of winter wheat farmland in growing seasons of 2013-2015, then the simulated results were compared against the field experimental data at agrometeorological station in Nanjing, China. Furthermore, separation of soil evaporation and plant transpiration were conducted. Results showed that the R² values between the simulated and measured values reached 0.83 during 2013-2014 and 0.74 during 2014-2015. When the diurnal variation of leaf clumping index and root vertical distribution during the growth stage of the winter wheat was adopted in BEPS, BEPS performed better after modification, with the model efficiency (ME) value increasing from 0.74 to 0.8, the root mean square error (RMSE) value decreasing from 1. 50 mm to 1.05 mm, and the mean bias error (MBE) value decreasing from 0.50 mm to 0.35 mm, respectively. The ratio of soil evaporation to ET changed with growth progress during the growing stage, and the growing seasonaveraged ratios of soil evaporation to ET were 34% and 29% for the two growth seasons, respectively. The BEPS model explained 78% variations of soil moisture in the depth of 0-40 cm, and responded quickly to the change of precipitation. Our study indicates that the improved ecological model can be used to simulate the changes of ET and soil water content in winter wheat farmland, and estimate the ratio between transpiration and evaporation that are difficult to distinguish in crop canopy. Our results will provide a basis for evaluating regional ET and agricultural watersaving management under climate change in the future.