In Ethiopia, there are 32.85 million sheep, more than 99% of which are indigenous. However, the productivity of local sheep under traditional production system is low with high mortality of lambs. Between sheep breed differences in performance do exist and have begun to attract some research attention in recent years. Yet, little selection has been undergone for improved meat production and true breed potentials are not known and not all breeds have been sufficiently well characterized. There is a need to better acknowledge the unique features of these breeds, their products, their socio-cultural and religious uses, and their production systems. Multiple-visit, multiple-subject survey was conducted to characterize sheep production system, growth and reproductive performance, phenotypic features and genetic diversity study and generate information’ to design sheep breeding schemes. Purposive sampling method was employed to select study areas or districts and households engaged in sheep production. Pre-teested semi-structure questionnaire was administered for data collection. Data collection was carried out for two years from September, 2017 to October, 2019 G.C. Households under investigation have a small patch of land. Mean ±SD values of land holding per household of Burie, Mandura and Quara were 1.03±0.89 ha, 2.66±1.61 ha and 11.46±5.76 ha, respectively, resulting in insufficient pasture and feed shortages. Feed shortage, disease and parasites and lack of input were prioritized as major challenges of sheep production. The growth and reproductive performances traits of Washera and Gumuz sheep were comparable to other Ethiopian and African sheep. Breed type affected live weight significantly (P<0.001). Average age at first lambing and lambing interval were also affected by breed significantly. Body weight recorded for Washera, Gumuz and Rutana Sheep were 27.5, 26.89 and 38.71 kg, respectively. Chest girth and body length were best body weight predictors for Gumuz female sheep at p<0.001 level of significance. Tail length, ear length and body length were best body weight predictors for Gumuz male sheep at p<0.001 level of significance. For Washera female sheep, tail length, chest girth, height at withers and body length were best body weight predictors at p<0.001 level of significance. Tail length, chest girth and head length were best body weight predictors for Washera male sheep at p<0.001 level of significance. Higher mean values of morph metric traits were recorded canonical bone length. The variations in phenotypic features need to be supplemented by molecular diversity study. Clustering analysis using principal component (PC1) differentiates Rutana from Gumuz and Washera sheep supporting this variation. Polymorphic information content was 0.89, 0.90 and 0.88 for Gumuz, Rutana and Washera sheep, respectively. These values could be recognized as reasonably (moderately) informative to carry out genetic diversity study. Expected heterozygosity was found to be higher than observed heterozygosity attributed to the discrepancy to forces such as inbreeding. Observed heterozygosity ranged from 0.33 in Gumuz to 0.34 in Rutana and Washera sheep. Expected heterozygosity ranged from 0.37 in Rutana to 0.38 in Gumuz and Washera sheep. Higher inbreeding coefficient (0.12) was recorded for Gumuz sheep compared to 0.09 of Rutana and Washera sheep. Our whole-genome SNP analysis revealed high variation within populations. These within-breed variabilities were used as input for designing a selective breeding program. Phenotypically, two highly rated selection criteria (body size and litter size) were preferred for Gumuz and Washera ewe. Improving 25 kg sale weight (at yearling age) and the number of lambs qualifying for the market and home consumption were the proposed breeding objectives or goals. Genetic gain predicted for six- month weight of Gumuz and Washera sheep were 0.46 and 0.60 kg of live weight, respectively. Genetic gain predicted for yearling weight of Gumuz and Washera sheep were 0.58 and 0.66 kg of live weight, respectively. A sheep breeding program with 10% selection proportion and one year ram use for breeding was recommended giving special emphasis to yearling weight.