In this study, ten cotton varieties were setup in a partial diallel crosses. These varieties were G. 45 (P₁), G. 70 (P₂), G. 86 (P₄), G. 88 (P₅), G. 89 (P₆); G. 90 (P₇), G. 77 (P₁₀) were Egyptian cotton varieties, Pima 62 (P₃) and Pima S₆ (P₈) were American varieties, as well as, Kar. (P₉) was Russian variety, all these varieties belonging to the species Gossypium barbadense L.                 In the growing season of 2006, the ten parental varieties were planted and crossed in a half diallel mating design to obtain 45 F₁ single crosses. The all genetic materials used in this investigation included ten parental varieties and their 45 F₁ single crosses were evaluated in the 2007 growing season at Sakha Agriculture Research Station at a randomized complete blocks design with three replications. Gene action, general and specific combining abilities and heritability in broad and narrow senses were estimated for boll weight, seed cotton yield per plant, lint cotton yield per plant, lint percentage, number of bolls per plant, seed index, lint index and number of seeds per boll. The results showed that highly significant differences among all evaluated genotypes for all the studied yield and yield component traits. The mean performances of genotypes revealed that G. 86 (P₄) had the highest mean values for boll weight, seed cotton yield per plant, lint cotton yield per plant and lint percentage. While, the Kar. (P₉) was the best and had the highest means for seed index, lint index and number of seeds per boll. The results also indicated that, the mean performances of most the 45 crosses were better than their both parents. Such as, the crosses G. 90 x Kar., Pima S₆ x G. 77, G. 88 x Kar., G. 45 x Kar., Pima S₆ x G. 77, G. 70 x Pima S₆, Pima 62 x G. 86 and G. 90 x Kar. for boll weight, seed cotton yield per plant, lint cotton yield per plant, lint percentage, number of seeds per boll, respectively. The Kar. variety (P₉) was the best combiner and desirable for all the studied traits with except G. 77 (P₁₀) was the best combiner for lint percentage, and the results also, cleared that most of studied crosses had highly significant and positive (desirable) specific combining ability effects. The results revealed that the magnitudes of non additive genetic variance including dominance (s²D). Which were positive and larger than those of additive genetic variance (s²A) for all the studied traits. This finding indicated the importance of dominance genetic variance in the inheritance of all the studied traits. These results cleared that could be utilized these superior crosses in breeding programes for improve yield and its component traits or a lines as through the selection in segregation generation.