To reveal the linkage between κ-CN genotypes resulted from PCR-RFLP analysis and milk yield and composition traits, stepwise analysis was used for determining the best regression equation of those phenotypic traits on detected genotypes. 80 blood samples were used for genetic characterization of studied buffaloes, collected from five different geographical locations of Egyptian provinces with total of 340 dairy production records for 59 lactating buffaloes. Milk components profile was analyzed by MilkoScan device. Results of stepwise regression analysis showed that AcuI genotypes were significantly linked to most studied traits such as, TMY, FAT%, TS%, Ash%, and Humidity% with highly statistical. In some cases, the covariance between the independent variables (AcuI, HpyCH4IV, or compacted genotypes) represented the best regression equation, as with DMY and lactose%. Least-square means for studied traits showed that, heterozygous animals were significantly superior to homozygous animals, especially within AcuI or compacted genotyped animals. In other words, they could be considered the most superior individuals producing better milk with distinct fat%, TS%, SNF%, and Ash% traits. On contrary, homozygous animals had higher lactose percentage and humidity traits. Consequently, polymorphisms of κ-CN gene should be involved within modern selection programs as potential candidates associated with dairy performance traits referred to as gene assisted selection (GAS) that permits to select animals at an early age for breeding programs.