A complete set of diallel crosses among seven durum wheat lines were used to study the genetics of yield and its components. Eight traits including days to heading, harvest index, grain number.spike-1, grain weight.spike-1, grain number.plant-1, plant height and grain yield.plant-1 were measured and recorded. Significant (P<0.01) differences were observed among genotypes for all the studied traits. Assumptions of Jinks-Hyman method were valid for harvest index, grain number.spike-1, grain weight.spike-1 and grain yield.plant-1. However, for the remaining traits, when one parent was excluded, the assumptions fulfilled. Based on the estimates of average degree of dominance and result of graphical analysis, the gene action for days to heading, was partial dominance, while grain weight.spike-1, plant height, fertile tiller number, grain number.spike-1, grain number.plant-1, grain yield.plant-1 and harvest index were controlled by over dominance gene action. Hayman analysis of variance indicated that additive (a) effects were significant for all of the studied traits, and dominance (b) effects significant for all of traits except day to heading. Maternal effects (c) were also significant for all of traits except plant height and fertile tiller number. Broadsense heritability estimates ranged from 0.93 (for grain number.spike-1 and grain weight.spike-1) to 0.70 (for plant height). Narrow-sense heritability estimates ranged from 0.70 (for days to heading) to 0.30 (for fertile tiller number).The biplot indicated that entry Prion-1 had the highest GCA effect for grain weight.plant-1, fertile tiller number, grain number.plant-1, grain yield.plant-1 and harvest index. However, it had the lowest GCA for plant height and days to heading. Therefore, this genotype can be considered as suitable parent in durum wheat breeding programs.