The present investigation on "Inheritance of grain yield and its components in Triticum aestivum L." was undertaken in three crosses, each with the set of six generations viz., P1, P2, F1, F2, B1 and B2.The present investigation revealed that cross NI-5439 x PHR-1009 can provide better opportunities for improvement through selection procedure than rest of the crosses i.e. NIAW-917 x VL-852 and HD-2189 x FLW-8. Though significant dominance and dominance x dominance gene components were observed for many characters in the three crosses studied, it cannot be exploited because of presence of duplicate type of epistasis. In such a situation maximum gain could be attained by maintaining considerable heterozygosity through mating of selected plants in early segregating generations or some sort of Recurrent Selection, Diallel Selective Mating System (Jensen, 1970) or Biparental Mating (Joshi and Dhawan, 1966) should be practiced. This would certainly enhance possibility that vvarious recombinations may result in the accumulation of favourable genes in ultimate homozygous lines.few cycles of recurrent selection followed by pedigree breeding approach can be suggested to improve the Yield.