To adjust breeding programs for local, commercial, and fancy breeds, and to implement molecular (marker-assisted) breeding, a proper comprehension of phenotypic and genotypic variation is a sine qua non for breeding progress in animal production. Here, we investigated an evolutionary subdivision of domestic chickens based on their phenotypic and genotypic variability using a wide sample of 49 different breeds/populations. These represent a significant proportion of the global chicken gene pool and all major purposes of breed use (according to their traditional classification model), with many of them being characterized by a synthetic genetic structure and notable admixture. We assessed their phenotypic variability in terms of body weight, body measurements, and egg production. From this, we proposed a phenotypic clustering model (PCM) including six evolutionary lineages of breed formation: egg-type, meat-type, dual purpose (egg-meat and meat-egg), game, fancy, and Bantam. Estimation of genotypic variability was carried out using the analysis of five SNPs, i.e., at the level of genomic variation at the NCAPG-LCORL locus. Based on these data, two generally similar genotypic clustering models (GCM1 and GCM2) were inferred that also had several overlaps with PCM. Further research for SNPs associated with economically important traits can be instrumental in marker-assisted breeding programs.