During three consecutive years (2002 to 2004), common bean was grown to investigate the effects of the regulated deficit irrigation (RDI) on the yield and yield parameters under semi-arid conditions. Field experiments were conducted on a clay soil. The growing season of common bean was divided into two phases: (1) Vegetative stage (V); from seed germination to the beginning of flowering and (2) reproductive stage (R); from the flowering to the last fruit harvesting. The statistical design was a split-plot with three replications, where the growth stage was the main factor of variation and the irrigation was the secondary factor. The irrigation treatments consisted of all possible combinations of full irrigation (T-1 and T-5 (V-100 and R-100: 100% of irrigation water (IW)/cumulative pan evaporation (CPE)) or limited irrigation (T-2: V-75-R-100, T-3: V-50-R-100, T-4: V-25-R-100, T-6: V-100-R-75, T-7: V-100-R-50 and T-8: V-100-R-25) in two phases. Fresh bean yield and pod weight (PWt), pod length (PL), pod width (PWh) and number of seed per pod (NSPP) were measured. Yields of T-2, T-3, T-4, T-6, T-7 and T-8 were determined as 27.0, 35.0, 41.0, 4.0, 12.0 and 21.0% lower than the yields obtained from the control (18.36 and 18.40 t ha(-1)) treatments, respectively. Results demonstrated that, vegetative stage was the more sensitive than the reproductive stage to the deficit irrigation. The highest irrigation water use efficiency (IWUE) and water use efficiency (WUE) were found in T-6 and T-7 treatments as 2.58 and 2.66 kg m(-3), respectively.