The present investigation was conducted at Arab El-Awammer Agricultural Research Station, Assiut, Egypt during 2012, 2013 and 2014 summer seasons. Two cycles of pedigree selection method were applied on a segregating population of sesame in the F₃, F₄ and F₅ generations to improve seed yield. The soil of the experiment site is sandy calcareous. Highly significant differences among F₃ families and a sufficient genotypic coefficient of variability (GCV) were obtained for all studied traits. The coefficient of variability decreased from 49.88 % in the F₃ population to 22.68 % in the F₅ for seed yield/plant. Broad sense heritability estimates were high for most traits and reached 93.32 % for seed yield/plant after two cycles of selection. The observed gain in seed yield/plant was significant (P<0.01) and accounted for 80.27 % from the unselected bulk sample. All of the families in the F₅ generation showed significant increase in seed yield/plant. It could be concluded that pedigree selection was an efficient method to improve seed yield productivity in the studied population. Selection for productivity could be practiced in the early segregating generations and two cycles of selection were enough to detect the families with high productive. Otherwise, single trait selection accompanied with adverse effects on other traits, and selection index could be recommended to improve seed yield and other traits.