Reference evapotranspiration (ΕΤ₀) is a major component of the hydrological cycle. Its use is essential both for the hydrological rainfall–runoff assessment models and determination of water requirements in agricultural and forest ecosystems. This study investigates the performance of 24 different methods, which produce ET₀ or potential evapotranspiration estimates above a grass-covered ground in a Mediterranean forest environment in Greece and compares the derived results with those of the presumed most accurate and scientifically acceptable Penman–Monteith method (ET_P-M). Their performance was evaluated on a daily basis for a period of 17 years, using 17 different statistical parameters of goodness of fit. The results showed that some empirical methods could serve as suitable alternatives. More specifically, Copais (ET_COP), Hargreaves original (ET_HAR), and Valiantzas2 (ET_VA2) methods, exhibited very good values of the model efficiency index, EF (0.934, 0.932, and 0.917, respectively) and the index of agreement, d (0.984, 0.982, and 0.977, respectively). Additionally, the differences of the estimated mean daily value against the respective ET_P-M value (rt index) for all methods had a range of −27.8% (Penman – ET_PEN) to +59.5% (Romanenko – ET_ROM), while Copais (ET_COP), Hargreaves–Samani modified1 (ET_HS1), and STU (ET_STU) yielded the best values (−0.06%, +0.06%, and 0.22%, respectively).