The present study was carried out during the three growing seasons 2011, 2012 and 2013 at Experimental Farm, Fac. of Agric., ZagazigUniversity. A half diallel crosses among eight yellow maize inbred lines i.e. Z12 (P1), Z15 (P2), Z167 (P3), Z147 (P4), Z40 (P5), Z56 (P6), Z58 (P7) and Z103 (P8) were evaluated under well-watered and water stress environments. The data were recorded for the days to 50% silking, anthesis silking interval, plant height, ear leaf area, leaf water content, ear diameter, ear length, number of rows per ear, number of kernels per row, 100-kernel weight,  grain yield (ard./fad.). The combined analyses of variance for combining ability revealed that mean squares for general combining ability (GCA) and specific combining ability (SCA) were highly significant across two environments for all studied characters, indicating the prevalence of additive and non-additive gene action in the gene expression of these characters, but the ratio of GCA/SCA variances were more than unity for days to 50% silking, plant height and number of rows per ear, indicating the major role of additive gene effects in controlling the genetic mechanism of these characters over water environments. In contrast, the ratio of variance GCA to variance SCA was blow one for anthesis silking interval, ear leaf area, leaf relative water content, ear diameter, ear length, number of rows per ear,  number of kernels per row, 100-kernel weight and grain yield. This emphasized that, non-additive gene action was the prevailed type in controlling these characters. Narrow sense heritability estimates were high (> 50%) for days to 50% silking, plant height, ear diameter and number of rows per ear, moderate for ear leaf area (41.68%) and ear length (45.55%), and low (< 30%) for anthesis silking interval, leaf relative water content, number of kernels / row, 100-kernel weight and grain yield over two environments. Reduction % due to water stress valued, (32.83%) for grain yield (ard./fad.), (15.60%) for plant height, (14.99%) for ear length, (13.99%) for ear leaf area, (12.17%)  for 100-kernel weight, (10.63%) for number of kernels/row, (6.98%) for ear diameter, (6.64%) for leaf relative water content and (4.22%) for number of rows/ear. The best parental combiners were Z167(P3) and Z147(P4) for grain yield and most its components under optimum irrigation and water deficit. The crosses, (P3 x P6), (P2 x P3), (P4 x P7), (P4 x P8) and (P6 x P8) had the most desirable and highest values for mean performance, SCA effects, heterosis effects, droght toleranc index (DI) and stress tolerance index (STI) for grain yield, anthesis silking interval and other performance traits. The yellow maize crosses (P2 x P3) and (P4 x P8) were significantly outyieled the check varieties. Grain yield had positive and significant genotypic and phenotypic correlations with ear leaf area (0.443** and 0.355**), leaf relative water content (0.488** and 0.307**), ear diameter (0.691** and 0.546**), ear length (0.783** and 0.647**), number of rows per ear (0.291* and 0.237), number of kernels per row (0.486** and 0.451**), 100-kernels weight (0.659** and 0.543**) and drought susceptibility index (0.484** and 0.388**, respectively), but had negative correlations with days to 50% silking (-0.034 and 0.004) and anthesis silking interval (-0.572** and -0.491**, respectively). The results showed that ear length exhibited the largest direct effect on grain yield (0.340) followed by drought susceptibility index (0.251), leaf relative water content (0.231), ear leaf area (0.182), number of kernels per row (0.171), ear diameter (0.135) and number of rows per ear (0.104).