This study was conducted on the experimental farm of the New Valley Agricultural Research Station, Agricultural Research Center, Egypt, from 2008/2009 to and 2010/2011 seasons. The objective of this study was to estimate efficiency of pedigree, modified bulk and single seed descent methods in two durum wheat populations. The final evaluation of F5 generation was done during 2010/2011 season. The selected plants were sown in two experiments using randomized complete block design with four replications. Efficiency of each breeding methodology was evaluated on the basis of following parameters: mean performance, deviation between method of breeding and check, phenotypic and genotypic variances, phenotypic and genotypic coefficients of variability and broad sense heritability. Results indicated that, differences among genotypes, among selection methods and among selected families within all breeding selection methods were significant for all studied characters. Mean and actual response to selection possessed low values in pedigree method for days to heading, and in single seed descent and modified bulk methods for plant height. Whereas, mean and deviation between method of breeding and check possessed high values in pedigree and single seed descent methods for spikes per plant, and in pedigree and modified bulk methods for kernels per spike, kernel weight and grain yield per plant. On the other hand, broad sense heritability exhibited high values in pedigree method for plant height, kernels per spike, kernel weight and grain yield per plant and in modified bulk method for spikes per plant, and in single seed descent method for days to heading, plant height, spikes per plant and kernel weight. These results refer that the three methods are effective in improving shortness, spikes per plant, kernels per spike and 100- kernel weight, while, pedigree and modified bulk methods are effective in improving earliness and grain yield per plant.