The use of FAO-56 reference evapotranspiration (ET_o) for irrigation scheduling has been recommended worldwide because it provides the best results under a wide range of climatic conditions. Crop ET (ET_c) is calculated from ET_o using a stage-dependent crop coefficient (K_c) which depends on canopy height, crop-soil surface resistance and albedo. There is lack of information on K_c for cotton grown under deficit irrigation in humid regions in the southeastern USA. The objective of this study was to evaluate the K_c FAO-56 procedure under different deficit drip irrigation treatments for cotton grown in a humid climate. An experiment was conducted in 2005 with cotton grown in three automated rainout shelters located in Georgia. When the actual soil water content in the effective root zone dropped below a specific threshold of the available water content (AWC), irrigation was applied until the soil water reached 100% of AWC. The irrigation treatments were 40%, 60% and 90% irrigation thresholds (IT). Observed daily ET_c was estimated from daily soil moisture readings. The length of the initial, development and mid stages was the same between the different treatments. However, the length of late-stage for the 60% and 40% IT was 3 and 5 days less than that for the 90% IT, respectively. The initial and end K_c for the 90% IT was greater than for the 60% IT, which in turn was greater than the 40% IT. For the mid stage, the 60% and 90% IT had a similar value for K_c, which was higher than the 40% IT. The FAO-56 proved to be an accurate method for estimating cotton ET_c under deficit irrigation. The values for the mid and end K_c and the length of the different developmental stages of the 90% IT may be used in the humid southeastern USA for fully irrigated cotton.