The studying of gene action would be of great importance to plant breeders as it provides information about possible improvement of different yield, yield components and fiber properties traits. Hence, this investigation has been done to partition the genetic variance to its components through studies on different generations of two cotton crosses i.e. Giza 88 x  Pima S6 (Egyptian x American variety) and Giza 45 x Suvin (Egyptian x Indian variety) at Sakha Agricultural Station, during three successive growing seasons (2008-2010).     The results showed presence of significant differences among generations in the two crosses for all studied traits. These findings reflected the presence of heterotic effects and the higher frequency of dominant genes controlling these traits. Also, F₃^,s generation showed superiority for most studied traits compared with the F₂^,s generation values in two crosses. These results indicated that the parents Suvin, Pima S₆, Giza 88 and Giza 45 could had transmitted their performances to their offspring, hence could be utilized these parents in cotton breeding program for improving these traits. Highly significant positive heterosis was observed relative to mid-parents for most studied traits. In addition, heterosis relative to the better parent was significantly positive for all studied traits in the two crosses except for seed index, 2.5% span length, fiber fineness and fiber strength in cross I. As well as, seed index only in cross II.  Moreover, positive highly significant heterosis relative to better-parent were obtained for number of opening bolls / plant and lint percentage in cross I  and for all studied traits except, boll weight, seed index , 2..5% span length and fiber strength in cross II. Highly significant positive inbreeding depression values were observed in F2 and F₃ generations for most of yield and fiber quality traits with respect to the studied two crosses. Over dominance appeared to be controlling most studied traits in F₁ hybrids and F₂ generations in the two crosses and the other remaining traits were controlled by partial dominance. The mean effect of F₂ performance (m) was highly significant for all studied traits in the two crosses. Also, the additive gene effects (d) were significant or highly significant positive for all studied traits except uniformity ratio in cross II. While, Dominance effects (h) were positive and highly significant for number of opening bolls/ plant, lint cotton yield /plant and lint percentage in two crosses and for 2.5% span length and uniformity ratio in cross II. Therefore, the presence of both additive and non-additive gene action for most studied traits with some exceptions for certain crosses, indicated that selection procedures based of the accumulation of additive effects should be successful in improving these traits.         Finally, all types of gene action effects (d, h and epistasis) were highly significant or significant, but additive x additive component (i) epistatic effect was significant and higher in magnitude compared to other components. Therefore it could be concluded that the gene action played a major role in the inheritance of these traits.