This study evaluates the performance of 50 empirical models for estimating potential evapotranspiration (PET) using daily meteorological data from a semi-arid region in Tamil Nadu, India. These temperature- and humidity-based models were compared against the FAO56 Penman-Monteith (PM) model, a globally accepted benchmark. Statistical indices such as the coefficient of determination (R2), mean absolute error, standard error estimate, and long-term average ratio (rt) were employed to assess model accuracy and reliability. The results revealed that certain models-specifically Althoff (Water 11:2272, 2019), Pereira (Agric Water Manag 66:251-257), and Samani (J Irrig Drain Eng 126:265-267, 2000)-exhibited strong agreement with the FAO56 PM model, offering a robust balance between accuracy and simplicity. The Althoff et al. model achieved the highest ranking based on standardized performance indices. This study is significant in identifying cost-effective alternatives for PET estimation in data-scarce, semi-arid regions, which is essential for irrigation planning and water resource management. However, a key limitation of this research is its exclusion of radiation and wind-based models, which may provide enhanced accuracy under certain conditions. Future research should explore the integration and calibration of multiple climatic parameters to improve PET estimation further.