Seven CMS and seven restorer lines of grain sorghum [(Sorghum bicolor (L.) Moench)], their forty nine single cross hybrids and the check hybrid Shandaweel-1 were evaluated in two field experiments in 2007 season. The first experiment was irrigated normally as recommended, while in the second experiment, one irrigation was skipped after every normal irrigation (water stress). Combined analysis over the two irrigation treatments showed highly significant differences between the two irrigation treatments and among genotypes. Mean square due to the interaction of genotypes x irrigation treatments was also highly significant, indicating that genotypes responded differently to irrigation treatments. Furthermore, mean squares due crosses (C), parents (P), P vs. C, females (F), males (M), F x M and their interaction with irrigation treatments were significant for all studied traits except days to 50% flowering. Line x tester analysis indicated that the linesICSA-364and ICSR-66 showed the highest significant GCA effects for grain yield/ plant. The GCA effect for days to 50% flowering of lines ICSA- 363, ICSA-572 and ICSR- 102 and were negative and significant over the two irrigation treatments and they considered good combiners for earliness.  Better parent heterosis was generally manifested for plant height panicle length and width and grain yield/plant. The cross (ICS.A-610 x ICS.R-31) had highest positive significant heterosis for grain yield (66.97%).Crosses (ICSA-364x ICSR- 66), (ICSA-364x ICSR- 102) and (ICSA-490x ICSR- 66) had higher grain yield than the check (Shandaweel-1) and it should be produced commercially after tested on a large scale.