Bulk selection method was used among six faba bean crosses during F₃ and F₄ generations under drought and/or non-drought conditions.Four methods (M₁, was grown under drought in F₃ and F₄, M₂, was grown under drought in F₃ and non- drought in F₄, M₃, was grown under non-drought in F₃ and drought in F₄ and M₄, was grown under non-drought in F₃ and F₄) were formed from each cross in F₅and evaluatedin two open field experiments (under drought and non-drought conditions). The experiments were planted in split plotdesign with crosses in main plots and selection methods in sup plots. Highly significant differences among crosses and methods and their interactions were reported by combined analysis of both environments.Drought harmful faba bean traits and seed yield decreased from 53.5 to 74.4%. Under drought stress, M₁ selection method produced the highest seed yield andcrosses of Hassawi2 by Luz and by TW had the highest seed yield/plant 15.4 and 13.3g, respectively. On the other side M4 was the best selection method for non-stress conditionsand TW x Hassawi2, TW x Hassawi3 and Luz x Hassawi3 recorded highest seed yield 41.7, 37.9 and 37.3g, respectively. High values of PCV, GCV for M₄under non-drought suggested that bulk selection under favorite conditions increased the phenotypic and genotypic variance, particularly under same conditions. heritabilityvalues were higher for each selection method under conditions that breed for it indicating that, selected high yields under specific condition may be combined with sensitive alleles to other condition.Plants exposed to different conditions of drought and non-stress during the previous generations (M₂ and M₃) were had less response to environments changes with low seed yield.