Evaporation and Evapotranspiration Process are the major components of the hydrologic cycle which play an important role in agricultural studies such as design of irrigation and drainage systems, and irrigation scheduling. Evapotranspiration is a complex non-linear phenomenon which depends on several climatologic factors. It can be measured directly by high-cost micrometeorological techniques. Hence, many mathematical models and empirical equations were developed to estimate this phenomenon. One conventional method to estimate reference crop evapotranspiration (ET0) is converting the class A pan evaporation (EPan) into ETo by using a pan coefficient (KPan) according to following this equation. ETo = Kpan  Epan Another alternative method to estimate ETo is the application of mathematical models like artificial neural networks (ANNs). ANNs are mathematical models whose architecture has been inspired by biological neural networks. ANNs are very appropriate for the modeling of nonlinear processes, i.e. the case of ETo.Kpan is the important factor for computation of ETo from Epan, There for several empirical equations purposed to determine KPan, using wind speed, relative humidity and fetch length conditions by many researchers. The main objective of this study was to comparison between ability of ANNs and empirical equations for estimation daily ET0 from Epan. For this object Daily measured weather data for a 16 year (from 1992 to 2007) period were obtained from the Shiraz synoptic station (latitude 29o 36` N, longitude 52o 32` E, elevation 1480 m) that located in Fars province of Iran. The climate in the study area is semi-arid with an average annual rainfall of 346 mm year

Keywords