ABSTRACT: Epistasis, additive and dominance components of genetic variation for yield and some yield related traits were assessed through triple test cross technique in bread wheat. Two genotypes of bread wheat, Gemmeiza 11 and Line 1, and their F1 progeny were used as a testers to crossed with twenty random selected F2 plants from the previous cross (Gemmeiza11 × Line 1 ) to produce L1i (P1 × F2i), L2i (P2 × F2i) and L3i (F1i × F2i) respectively. The sixty families (L1 (20) + L2 (20) + L3 (20)) were sown at Experimental Farm of Gemmeiza Agriculture Research Station to study gene action, predicting of new recombinant lines and genetic correlation for some quantitative traits. The mean squares of the analysis of variance revealed significant differences among triple test cross (T.T.C) families for all the traits studied indicating that L1i, L2i and L3i families were significantly different from each other, providing evidence for adequate amount of genetic variability. Epistasis was found an important part of genetic variation for all traits studied. Partitioning of total epistasis into (i) type and (j +L) types of epistasis revealed that (i) type of epistasis (additive x additive) was found to be significant and highly significant for most traits studied except plant height, main spike length, number of spikelets per main spike and kernels per spike. However, (J+ L) additive × dominance, dominance × dominance, types were also highly significant for most traits studied except number of days to heading. The additive genetic variances (D) was found to be much larger in magnitudes than the dominance genetic variance (H) for number of days to maturity, number of spikes per plant, number of kernels per main spike, grain yield per plant and 1000- kernels weight and that resulted in (H/D)½ to be less than one. The results showed that the F^ value (covariance of sums / differences) was found to be significant and positive or negative for number of days to maturity and flag leaf area revealing that the dominance was unidirectional among parents. The prediction results revealed that it could be feasible to predict as early as possible for transgressive segregants and the highest proportions of recombinants which outperform parental range for number of spikes per plant, followed by number of days to maturity, grain yield per plant, flag leaf area and plant height. Thus the breeder should give a great emphasis to the promising cross are the most frequent ones and having high values for new recombinants for yield, therefore, the breeder should pay great emphasis for considering these promising cross in wheat breeding program. Results of genetic correlation generally revealed presence of significant additive, dominance and epistatic genetic correlations among some traits suggesting common genetic pool, pleiotropy or linkage. Since additive genes is a fixable type, therefore, selection based on such type may indicate that indirect selection via, main spike length, number of kernels per main spike, number of spikes per plant and 1000-kernels weight would be effective in improving grain yield and enhance its importance as selection criteria. Meanwhile, the great part of traits showed non-significant genetic correlations and confirmed that T.T.C mating system was useful in break up undesirable linkage to obtain new recombinant lines.