This study was carried out to determine the crop water stress index (CWSI) for drip irrigated cotton grown on a heavy clay texture soil (Palexerollic Chromoxerert) under semi-arid climatic condition of East Mediterranean region for three years (2005 to 2007) in Adana, Turkey. Four irrigation treatments designated as full (I-100) with no water stress and slight (DI70), moderate (DI50) and strong water stress (continuous stress, dry land) (DI00) were tested. The treatments of DI70 and DI50 received water amount of 70 and 50% of the control treatment and the DI00 was not irrigated except for germination water given at the beginning of the growing season. Irrigation was initiated when leaf water potential (LWP) reached to -15 bar for full (I-100), -17 bar for DI70 and -20 bar for DI50 irrigation treatments. After first irrigation, all the treatments were irrigated at one week interval. The deficit irrigation affected, the irrigation water use, seed cotton yield, dry matter and some yield components such as plant height and number of boll per plant of cotton. Average values of water use, seed cotton yield, dry matter and water use efficiency of full irrigated cotton were 578 mm, 3.28 tha(-1), 13.44 tha(-1) and 0.59 kg m(-3), respectively. CWSI values were calculated from the measurements of canopy temperatures by infrared thermometer (IRT), ambient air temperatures and vapor pressure deficit values for all the irrigated treatments. A non-water stressed baseline (lower baseline) equation for cotton was developed using canopy temperature measured from full irrigated plots as, T-c - T-a = -1.7543VPD + 1.56 ; R-2 = 0.5327 and the non-transpiring baseline (upper baseline) equation was built using canopy temperature data taken from continuous stress plots as, T-c - T-a = -0.0217VPD + 3.2191. The trends in CWSI values were consistent with the varying soil water content due to the deficit irrigation programs. The relationships between mean CWSI and plant parameters considered in this study were linear except for irrigation water amount. Both dry matter and seed cotton yield decreased with increased soil water deficit. Seed cotton yield (SY) and seasonal mean CWSI values relationship were obtained as, SY = -2.3552CWSI + 3.5657; R-2 = 0.499. This relationship can be used to predict the seed cotton yield. The results suggest that the cotton crop for this particular climate and soil conditions, should be irrigated when CWSI approaches 0.36. The CWSI approach, according to results of this study, can be accepted as a useful tool to schedule irrigations for cotton.