Banana (Musa spp.) is an important crop in the economies of many developing countries. In the north of Argentina, a subtropical region, banana plants grow in a suboptimal environment that limits yield because only one harvest per year is achieved. The objective of this work was to characterize the agroecological aptitude of Formosa, Argentina, for banana production through the behavior of three varieties of international use: Williams, Jaffa and Grand Naine, evaluated over five consecutive years. The three-way data analysis technique called Multiple Factor Analysis (MFA) was used for evaluating the varieties’ performances across cycles of production. The results allowed for inferring the existence of a genotype x environment interaction (GEI), corroborated by two-way factorial ANOVA. In order to determine how this suboptimal environment affected the development of each genotype of this perennial crop, Dual Multiple Factor Analysis (DMFA) was applied to jointly analyze the correlation structure between the traits that contributed to the performance of each variety in each year. The correlation structures between variables were different in each population and varied between years. All traits showed great variation between the years and genotypes, with the fruit peel thickness being the most discrepant throughout the years. However, Formosa appeared as a promising subtropical agroecological environment to produce banana because the varieties’ performances were acceptable for large-scale production systems. In addition to evaluating the adequate aptitude for cultivating banana in Formosa considering the significant effect of the GEIs, this research made a methodological contribution by proposing the use of three-way data analysis in Agronomy Science via MFA and DMFA.