We propose an efficient and time-saving strategy for starter culture selection. Our approach is based on the accomplishment of 3 phases: (i) the selection of yeast strains dominating spontaneous fermentations, (ii) the selection among the dominant strains of those showing the best technological characteristics, and (iii) the final selection among good technological strains of those showing the desired qualitative traits. We applied this approach to wine fermentations, even though the same strategy has the potential to be employed for the selection of any type of starter culture. We isolated and identified yeast strains at the mid- and final stages of 6 spontaneous fermentations carried out in 3 different Spanish wineries. We identified all strains as Saccharomyces cerevisiae by restriction fragment length polymorphism of the ribosomal DNA internal transcribed spacer region, and subsequently distinguished each strain by analyzing the polymorphism of the inter-delta regions. Strains that were detected both at the mid- and final stages of the fermentation were considered dominant. Four dominant strains were finally selected and tested in pilot-scale fermentation, and their performance was compared with that of a commercial wine strain. All dominant strains showed good fitness and resulted suitable to be employed as starter cultures. One of the dominant strains isolated in this study is currently commercialized.