To detect types of gene action especially epistasis for studied traits, six populations; P1, P2, F1, F2, BC1 and BC2 in two intra-specific crosses of canola (N.A.37 × Serw 4 and N.A.14 × Serw 6) were evaluated in a field trial during 2019/2020 season at Kafr-El-Hamam/Sharkia and Al-Arish/North Sinai Agricultural Research Stations, Agricultural Research Center, Egypt. Scaling and joint scaling tests revealed inadequacy of simple additive-dominance model in the inheritance of all studied traits in corresponding crosses at both locations. Significant negative heterosis and heterobeltiosis were found for days to first flower, plant height and first siliqua height, whereas significant positive were found for seed weight per plant and its components in the corresponding crosses at both locations. High to moderate values of narrow sense-heritability coupled with high (more than (20%) values of expected response from selection (as % of mean) were detected for days to the first flower in the 1st cross at Kafr-El-Hamam and in the 2nd cross at both locations, first siliqua height in the 1st cross at both locations and in the 2nd cross at Kafr-El-Hamam, 1000-seed weight in the 2nd cross at Al-Arish and seeds weight per plant in the 1st cross at both locations. The major role of dominance gene effects as the ratio ((H/D)0.5>1) along with the duplicate epistasis was detected in the inheritance of most studied traits in the corresponding crosses at both locations. Therefore, it is advisable to delay selection to advanced generation for improving most studied traits in most cases.