The aim of this study was to verify the adaptability and stability of soybean cultivars with regards to yield and oil content. Data of soybean yield and oil content were used from experiments set up in six environments in the 2011/12 and 2012/13 crop seasons in the municipalities of Patos de Minas, Uberaba, Lavras, and São Gotardo, Minas Gerais, Brazil, testing 36 commercial soybean cultivars of both conventional and transgenic varieties. The Wricke method and GGE biplot analysis were used to evaluate adaptability and stability of these cultivars. Large variations were observed in grain yield in relation to the different environments studied, showing that these materials are adaptable. The cultivars exhibited significant differences in oil content. The cultivars BRSGO204 (Goiânia) and BRSMG (Garantia) exhibited the greatest average grain yield in the different environments studied, and the cultivar BRSMG 760 SRR had the greatest oil content among the cultivars evaluated. Ecovalence was adopted to identify the most stable cultivars, and the estimates were nearly uniform both for grain yield and oil content, showing a variation of 0.07 and 0.01%, respectively. The GGE biplot was efficient at identifying cultivars with high adaptability and phenotype stability.