The current investigation was conducted at Shandaweel Agric. Res. Station, Sohag governorate, Egypt during 2018/2019 to 2020/2021 seasons. Pedigree selection of grain yield plant-1 was practiced independently from the F2 to F4 generations under normal and late planting dates. Evaluation was under both planting dates in the F4 generation. The phenotypic variance was slightly higher than the genotypic variance, and reduced from the F2 to F4 generation. Broad-sense heritability was 83.47 and 86.73% under normal planting compared to 88.22 and 89.03% under late planting after the first and second cycle of selection, respectively. The realized heritability was 37.75 and 40.63% under normal planting compared to 57.75 and 83.48% under late planting after cycle 1 and 2, respectively. The average observed gain of grain yield plant-1 after two cycles of pedigree selection was 12.59 and 25.33% from bulk sample and 7.45 and 2.69% from the better parent for normal planting selections, while it was 22.05 and 47.18% from bulk sample and 16.48 and 20.59% from the better parent for late planting selections, when the selected families were evaluated under normal and late planting conditions, respectively. The antagonistic selection increased the mean and decreased the sensitivity compared to the synergistic selection either evaluation was under normal or late planting conditions. Based on the path-coefficient analysis, number of spikes plant-1 had the greatest positive direct effect on grain yield plant-1 followed by number of kernels spikes-1 and 100-kernel weight in the base population and cycle two of selection under both planting dates.