The overall aim of this dissertation was to develop a community-based breeding program for sustainable use of Farta sheep genetic resources in northwest Ethiopia. The specific objectives were: 1) identifying sheep breeding practices and trait preferences; 2) evaluating the growth performance, lamb survival and economic benefits of Farta and their crosses with Washera sheep; 3) estimating economic values (EV) of traits; 4) estimating the genetic and phenotypic parameters; 5) evaluating the efficiencies of village-based selection schemes. Methodologically, sheep breeding practices and trait preferences of smallholder farmers were identified using survey and choice experiment approaches. For both approaches, data were collected from 370 farmers in three districts (Estie = 117, Farta = 125 and Lay Gayient = 128). The choice data were analyzed using the generalized multinomial logit model, while the weighted indices were calculated to describe the rank data. Subsequently, the growth performance and lamb survival of Farta and its cross with Washera sheep were compared using data collected from 4318 sheep (3195 crossbreds with 50% Washera, 115 with 75% Washera and 1008 pure Farta). The growth and survival data were analyzed using the general linear and Cox regression models, respectively. Farmers’ perceptions about the Washera-Farta crossbreds were examined using qualitative data generated through interviewing 53 farmers among those who participated in the crossbreeding program. In addition, economic valuation based on hedonic price modeling was conducted using data collected from 216 Washera-Farta crossbreds and Farta sheep traded in the local market. The growth and lamb survival data collected from the Farta sheep and Washera-Farta crossbreds were also used to derive EV for various traits and estimate the genetic and phenotypic parameters for growth and pre-weaning lamb survival traits. The five models evaluated for genetic and phenotypic parameters estimation were: 1) a model with random effect of direct animal genetics; 2) direct animal genetics with the permanent environment; 3) direct animal and maternal genetics; 4) a model with all the three random effects; 5) logit transformed form of model 4 (only for pre-weaning lamb survival). Lastly, the biological and economic efficiencies of three selection schemes were evaluated using the findings obtained in specific objectives 1 to 4 as input parameters. The schemes were one-stage selection at 6 months (I) or yearling age (II) and two-stage selection with the first at 6 months using a 50% selection proportion and the second at the yearling age (III). All schemes were evaluated with three levels of ram selection proportions (P = 0.05, 0.1 and 0.2) and four flock sizes (600, 1200, 1800 and 2400 breeding ewes). The results of the questionnaire survey showed that the average Farta sheep flock size per farmer was about 10.21±4.79 sheep. Having the highest index value of 0.36, income generation was ranked as the primary reason for keeping sheep, followed by meat consumption (0.31) and manure production (0.16). The majority of the smallholder farmers (91%) have the experience of selecting breeding rams and ewes within their flock using diverse criteria. Given the highest index value of 0.34, body size was ranked as a primary ram and ewe selection criterion, followed by coat color. Furthermore, choice modeling results revealed that good tail type, large body size and brown coat color were the most preferred attributes for breeding rams. For breeding ewes, the highly preferred attributes were good mothering ability, brown coat color and large body size. Regarding the growth performance of Farta and its crosses with Washera sheep, the least-squares means (LSM) of birth weight were 2.83±0.06, 3.16± 0.06 and 3.24±0.08 kg for Farta, 50 and 75% Washera crossbreds, respectively, in which both crossbred genotypes were significantly heavier than Farta sheep (P<0.001). However, Farta sheep were significantly heavier in weights at weaning age and 6 months than both crossbred genotypes (P<0.001). The yearling weights among the three genotypes were not statistically different (P>0.05), with the LSM being 22.17±0.33, 21.92±0.31 and 23.25±0.87 kg for Farta, 50 and 75% Washera crossbreds, in the order mentioned. The proportions of lambs that survived up to a yearling age were 89, 82.5 and 70% for Farta, 50 and 75% Washera crossbreds, respectively. All the interviewed farmers preferred keeping Washera crossbreds to Farta sheep for reasons such as a relatively faster growth at an early age, a large body frame, good coat color and a high market price of the crossbreds. The hedonic price model revealed that breed type was among the various factors that had significant influences on price (P<0.001) in which the Washera-Farta crossbreds had about 12.7% higher price than Farta sheep. Concerning EV estimates for production traits, it was found that a unit genetic change for weight at weaning, six months of age, in yearlings and mature ewes had resulted in EV of US$ 1.46, 0.60, -0.04 and -0.24 per ewe per year, respectively. In addition, the EV for the number of lambs weaned per ewe bred, ewe fertility rate, lambing interval, post-weaning lamb survival and mature ewe survival were US$ 0.59, 0.37, 0.16, 0.53 and 0.86, in the order given. Analysis of the sensitivity of EV to fluctuations in the price of input and output parameters revealed that the marginal change in profit was more affected in most traits when the price of live weight was altered rather than feed cost. Comparisons of the various genetic models showed that model 2 for birth weight and model 3 for weaning weight, pre-weaning daily gain and six-month weight were found to be more appropriate for genetic parameters estimation. However, model 4 was conceptually more insightful to describe the variations in yearling weight and pre-weaning lamb survival. Accordingly, the direct heritability for birth weight, weaning weight, pre-weaning daily gain, six-month weight, yearling weight and pre-weaning lamb survival were 0.08±0.03, 0.20±0.05, 0.15±0.05, 0.24±0.06, 0.29±0.14 and 0.14±0.04, respectively. The maternal heritabilities were 0.29±0.05, 0.31±0.05, 0.40±0.05, 0.15±0.57 and 0.02±0.15 for weaning weight, pre-weaning daily gain, six-month weight, yearling weight and pre-weaning lamb survival, respectively. The genetic correlation between birth weight and pre-weaning lamb survival was negative (-0.36±0.22). For the other pairs of traits, genetic correlations ranging from 0.10±0.16 between yearling weight and preweaning lamb survival to 0.98±0.01 between weaning weight and pre-weaning daily gain were estimated. Finally, the analysis of the efficiencies of the three alternative selection schemes revealed that the highest annual selection response for each scheme was predicted when P = 0.05, while the lowest estimates appeared at P = 0.20. It was also observed that increasing flock size had little effect on the genetic progress but significantly reduced the rate of inbreeding across the studied schemes. The annual economic responses estimated for schemes I, II and III varied from US$ 0.393 to 0.591, 0.589 to 0.879 and 0.494 to 0.744, respectively. In other words, the annual economic gain in scheme II was higher by 34 and 16% than those predicted for schemes I and III, respectively. Operationally, scheme I is more feasible than other schemes but offers less economic benefit, while scheme II is relatively easier to implement than scheme III and provides higher annual economic gains than schemes I and III. Overall, implementing scheme II with P = 0.05 and flock size of 1200 ewes would be appropriate for the genetic improvement of Farta sheep through a community-based breeding approach. Given the crossbreds’ low survival rate and absence of growth superiority at the marketing and breeding ages, Farta sheep genetic improvement through Washera ram-based crossbreeding was not found to be effective under the existing low-input production system. However, considering the preference of farmers and the price advantage in the local market, Washera ram-based crossbreeding could still be an option to improve the livelihood of farmers in areas not addressed by the community-based selection program if the existing sheep management practice is improved.