Effective soil water management is essential for successful agricultural growth and efficient water use. Evapotranspiration (ET) estimation plays a crucial role in hydrologic modelling that simulates soil water movement. While the FAO Penman-Monteith (PM) model is a popular choice for estimating reference crop ET (ETo), it depends heavily on various climatic variables often unavailable in data-scarce regions. This study examines eight different ETo models to simulate soil water dynamics in the low-lying paddy fields of eastern India in an effort to address the challenge of limited data availability. Using the HYDRUS-1D model and a non-weighing type paddy lysimeter, the actual ET was estimated. The results revealed that while all ETo models exhibited similar trends in simulating soil moisture content (SMC), their accuracy varied significantly. The models namely Blaney-Criddle, Hargreaves-Samani, Ivanov, Dalton, and Penman deviated considerably from the observed SMC values. However, for the FAO-PM, Irmak, and Priestly-Taylor models, no significant differences were detected between the simulated and observed values of SMC. The Irmak and Priestly-Taylor models can account for 94% of the variability in predicting SMC compared to the FAO-PM model. This study highlights that ETo models that require less data inputs, such as Priestly-Taylor and Irmak models, can also produce reliable and accurate results and can be used in water management related studies in eastern India. These models ensure efficient use of water in agriculture and offer practical solutions for regions where meteorological data is scarce.