The present study was carried out during the four successive seasons of 2009/2010, 2010/2011, 2011/2012 and 2012/2013 at the Experimental Farm of Faculty of Agriculture, Sohag University, Egypt. Bread wheat populations (Triticum aestivum L.) in F3, F4, F5 and F6 generations of the (Sids 12 x HAAMA-14) and (Giza 168 x TRI 2592) crosses were used in this investigation. One hundred 3 families of each population, which underwent pedigree selection in the 3 basic material for the derived F6 families used in this study. The aim was to compare the effectiveness of late pedigree selection vs. early pedigree selection in developing high yielding genotypes of bread wheat. Forty families were selected in F4 using grain yield, 100-kernel weight and days to heading as selection criteria. Twenty families were selected in F5 using the same criteria. However, ten families were selected as promising in F6 using late and early selection for grain yield. Analysis of variance showed highly significant differences between F3 families and satisfactory genotypic coefficients of variation, indicating the presence of sufficient variability for direct and indirect selection. After three cycles of selection in the F3 families, the genotypic coefficients of variability rapidly decreased for all studied traits. Estimates of broad sense heritability were relatively high and ranged from 76.73% for no. of spikes/plant to 99.94% for no. of kernels/spike in population 1 and from 72.34% for no. of spikes/plant to 98.61% for no. of kernels/spike in population 2. The expected genetic advance as percent of F3 ranged from 19.70% for days to heading to 94.94% for number of spikes/plant in population 1 and ranged from 23.21% for plant height to 79.24% for spike length in population 2. The late pedigree selection increased grain yield after one cycle of selection in the population compared to the best parent and the bulk population by 12.39 and 25.44% in population1 and by 9.26 and 23.74% in population 2. Grain yield increased after three cycles of early selection of both populations compared to the best parent and the bulk population by (20.66 and 34.67%); (17.08and 32.59%), respectively. 100-kernel weight increased by (11.32 and 24.25%); (7.98 and 22.28%), respectively, in both populations compared to the best parent and the bulk population. The best two families No. 42 and 56 were isolated from population1 increased more than the best parent by 28.58 and 31.52% using selections criteria for grain yield, respectively. In population 2 the best two families No. 52 and 56 increased more than the best parent by 24.74 and 27.60% using selections criteria for grain yield, respectively. After three cycles of selection of wheat realized gains indicated that heading date was reduced by -9.84 and -10.86% compared to the best parent of both populations. The high grain yield families using different criterion obtained from this study could be used in developing new wheat lines and effective for breeding methodology in developing high yielding.