In Europe, there is a growing interest for the use of breeds that shed their wool naturally. Indeed the weak relative value of meat and wool and the increasing shearing costs lead breeders to favour breeds that shed naturally annually. By using an original selection strategy, a new wool-shedding sheep breed was created in two consecutive steps. Firstly, a new genotype was created by introgressing the gene pool from the Martinik Hair, a hairy sheep known to shed, into the Romane breed through 4 consecutive backcrossing generations. For those animals, ability to shed and shedding extent increased by 75.4 and 130.2%, respectively, at 7 months of age compared to the Romane breed. Then the introgressed animals were selected during 8 generations on shedding extent over the body. To do so, a population of 150 ewes and 10 rams was selected on shedding extent estimated breeding values (EBV) under an annual breeding system with lambing period in early March. Fleece shedding extent was measured at mid-June in both adult ewes, and all lambs i.e. at the age of 3.5 months in lambs and once a year in ewes. All rams were replaced each year while ewe replacement varied from 30 to 50%. After 8 generations of selection of introgressed animals, all ewes showed a complete or nearly complete shedding of their fleece at mid-June: mean shedding extent was 89.3±17.9% in ewes indicating that annual shearing was not required anymore. In young lambs, mean fleece shedding extent measured at 3.5 months of age was 61.6±35.6%. Heritability estimate of shedding extent was high (0.38±0.04) and a genetic gain of 3 σg was observed after 8 generations of selection in this new breed deriving from the French Romane breed which has a high production potential. Breeding performances as well as lamb growth performances of this new breed were similar to those of the French Romane breed. Our selection strategy based on an introgression process of the gene pool of fleece shedding from a hairy moulting sheep is an interesting strategy to lead to a new shedding breed and/or improving adaptive traits in breeds having a high production potential. Current investigations are in progress to identify causal mutations in wool-shedding.