The simultaneous evaluation of qualitative and quantitative traits is an excellent tool to estimate the genetic divergence among accessionsof germplasm banks. The present study aimedto characterize and estimatethe genetic divergence among soybean food type accessionsbased on Gower algorithm and identifies promising combination among accessionsevaluated. A total of 64 soybean genotypes were studied, which 60arefood type soybean accessions, twoarecommercial cultivars also for human consumption,and two are grain type commercial cultivars. Three qualitative and seven quantitativetraits were used to estimate the genetic divergence among genotypes. The experiment was carried out in the experimental areaof the Universidade Estadual de Londrina during 2015/2016season, using a randomized complete block design with four repetitions. It was observed the presence of genetic variability to all traits evaluated. Thegrainyield trait was the mostefficientin discriminating theevaluated accessions,while plant height showed lower contribution. The UPGMA method showed the highestcoefficient of cophenetic correlation allowing the formation of six different groups. The crossings AL06 × AL22 and AL07 × AL22 are promising for obtaining superior genotypes destined to the in naturaconsumption and production of soy milk and tofu. The combination AL13 × AL60 is considered favorable for obtaining genotypes destined to natto production. The use of the AL12 and AL13 accessions as parents is advantageous for obtaining genotypes for the production of kuroname.