Increasing of yield is the most important objective of plant breeders in breeding programs. However, yield is a complex trait and includes several quantitative components. Its inheritance is polygenic, hence studying and improving the associated components to yield would promote yield. The objective of this study was to study the inheritance of quantitative traits in bread wheat using generations mean analysis (GMA). Sardari cultivar and inbred line 14, with different characteristics were crossed to develop the F1 generation. The F1 plants were selfed and backcrossed to both parents to produce F2 and backcross generations, respectively. All generations (F1, F2, BC1 and BC2) along with relevant parents were planted in field experimental station, Faculty of agricultural Tehran University, using a randomized complete block design (RCBD) with three replications. Generation mean analysis was performed using scaling test which asseses all generations simultaneously. The generations mean squares of the following traits: spike weight, plant height, number of kernel per spike, kernel weight per spike, 1000-kernel weight, awn length, spike length, awn status, spikelet compactness, kernel color, and number of tillers, were significant. The minimum number of genes controlling these traits, were 2, 2, 2, 2, 5, 1, 2, 4, 2, 2, 2 and 1, respectively. The average of broadsense and narrow sense heritabilities were 62% and 33%, respectively. Gene effects including mean, additive, dominance, epistasis effects as well as additive ´ additive, additive ´ dominance and dominance ´ dominance were observed for all traits. For awn length and number of tillers, additive and dominance effects and for other traits except awn status and glume color, the epistasis effects (specially additive ´ dominance and dominance ´ dominance) were determined as the most important effects in controlling the heritability of these traits. For spike weight, spike length, kernel weight per spike, awn length, plant height, spikelet compactness, and kernel color additive effect was more important than dominant effect. However, for number of kernel per spike dominant effect was more important than additive effect.