Traditionally, selection schemes in French dairy goats were based on genealogical and phenotypic information. However, in 2018, genomic information was introduced to preselect males for progeny testing (PT+GS). This scheme could evolve to a full genomic selection scheme (GS), avoiding the progeny test step in the near future. The objective of this study was to compare the current genomic selection scheme (GS+PT) with a traditional progeny testing (PT) and a full genomic selection scheme (GS) in French Saanen dairy goats. A population, under selection for a production trait (h2=0.3 and repeatability=0.5) was simulated using a base of real genotypes from 761 proven males. These animals were genotyped with the Ilumina GoatSNP50 chip. After filtering, 46,810 SNP remained for the analyses. One thousand SNP were selected to be QTL. The three selection schemes (PT, GS+PT and GS) were modelled using stochastic simulations to create phenotype and genotype data files to be processed, as in the real life, by single step GBLUP. Results based on 20 replicates were obtained. Compared to the PT scheme, the annual genetic gain was 30 and 35% higher for the PT+GS and GS schemes, respectively. The rate of decrease in genetic variance was the same for the three schemes. Selection accuracy for the young male candidates in the PT scheme was significantly lower (0.26±0.03) than in the GS+PT (0.39±0.02) and in the GS (0.40±0.02). Similar results were observed for young females. However, accuracy values for male candidates were lower than those for young females. Overall, these results confirm the advantage of genomic selection in goat breeding context.