To determine the gene action, heritability and number of genes controlling important traits in rice, two Iranian rice cultivars, Deylamani and Sepidroud were crossed in 2006. After selfing of F1 plants and back crossing with two parents in 2007, F2, BC1 and BC2 were developed. Seeds of two parents (P1 and P2) together with seed of F1, F2, BC1 and BC2 generations were grown in randomized complete block design with three replications at Research Field of the University of Guilan, Rasht, Iran, in 2008 cropping season. Eleven important agronomic traits including plant height, panicle length, panicle number per plant, spikelet number per panicle, grain number per panicle, empty spikelet number per panicle, panicle fertility percent, 1000 grain weight, days to 50 % flowering and days to maturity and grain yield were measured. Analysis of variance showed significant differences between generations for all traits, except 1000-grain weight and grain number per panicle. Therefore, generation mean analysis was performed to determine gene action and heritability of the traits of interest. Results showed that additive gene action for days to 50% flowering and days to maturity was more important than the dominance gene action, but for other traits, the dominance gene action was more important, expect for grain yield that was controlled by both additive and dominance gene actions. Furthermore, additive, dominance and epistatic effects had important roles in the inheritance of plant height, number of empty spikelets per panicle and grain yield. Estimation of degree of dominance also showed that days to 50% flowering, days to maturity and grain yield were controlled by partial dominance, panicle length by complete dominance and other traits by over-dominance effects. Average broad-sense heritability ranged from 0.36 to 0.82 for panicle length and days to 50% flowering, respectively, on the other hand, the average narrow-sense heritability varied from 0.07 to 0.69 for plant height and days to 50% flowering, respectively. Average number of genes controlling the traits ranged from at least one gene for panicle number per plant, spikelet number per panicle, panicle fertility percent and number of empty spikelets per panicle to 24 genes for grain yield. In conclusion, hybrid development will be an appropriate approach to improve plant height, panicle number per plant, spikelet number per panicle and number of empty spikelets per panicle.. However, for other traits, recurrent selection in segregating populations can be used to increase the frequency of favorable alleles, and then by hybridization between the selected lines with high performance the dominance gene effects are utilized.