Understanding the responding mechanism of reference evapotranspiration (ETo) to large-scale atmospheric circulation patterns is of particular concern in arid and semi-arid environments, where water resources scarcity constrains agriculture. This study investigates the statistical relationships between the Arctic Oscillation (AO) and the annual and seasonal mean ETo values of 41 weather stations in Iran for 1966e2005. The connections between the AO index and annual ETo are also examined during the extreme phases of the AO (high and low phases). Significant correlations between annual ETo and the corresponding AO index are quite rare, and the differences between the ETo values during the extreme AO phases and the long-term average ETo values were found to be significant only at three out of the 41 study stations. The lag correlations indicated that the greatest positive (0.506) and negative (0.466) correlation coefficients were found for 2-season time lag. Furthermore, the mean time lag between the start of the AO events and the observed maximum effects on ETo at the study stations was about 1.61 seasons or 5 months.