In view of the narrow genetic base of popcorn, probably due to its evolution by selection from flint maize types alone, knowledge about genetic divergence is imperative for the formation of heterotic groups. Thus, our objective was to identify heterotic groups of popcorn lines; we did so by exploiting the representative genotype collection of the Active Popcorn Germplasm Bank of the State University of Northern Rio de Janeiro. Thirty-eight popcorn genotypes from different origins were analyzed by two methodologies to identify divergent groups. In the first method, the genotypic data were processed to determine the number of groups, based on Bayesian clustering algorithms, and two clustering methods (UPGMA and Ward), based on three genetic distance algorithms, weighted index, unweighted index, and an index of genetic distance or dissimilarity, proposed by Smouse and Peakall. The second methodology identified groups based on simultaneous use of morphoagronomic and molecular information and extracting the genetic distance matrix by the Gower algorithm, and later applying UPGMA and Ward clustering methods. The consistency of the clustering methods was evaluated by cophenetic correlation coefficients. The significance of these coefficients was examined by the Mantel test. There was significant genetic variability among corn popcorn accesses at morphological and molecular levels. There also ©FUNPEC-RP www.funpecrp.com.br Genetics and Molecular Research 17 (3): gmr18083 C. Vittorazzi et al 2 was agreement between multivariate clustering techniques, mainly when using genotypic data provided by microsatellite markers. heterotic groups were identified; these were formed mainly according to the origin of each genotype and/or geographic origin. We found that there is sufficient heterosis to develop new cultivars.