Cotton (Gossypium barbadense L.) is a crop of great economic importance, as the superior quality of its fiber is highly valued worldwide. The objective of this research was to evaluate the agromorphological diversity of cotton germplasm using both qualitative and quantitative descriptors. A combination of univariate and multivariate statistical analyses was applied to identify promising accessions. The coefficients of variation (GCV and PCV) and high heritability estimates observed for descriptors such as the number of sympodial branches, fiber weight, and number of seeds per capsule confirm genetic control, ensuring the effectiveness of selection in future breeding programs. Correlation analysis revealed a strong positive relationship between lint yield and number of bolls per plant (0.893). Furthermore, principal component analysis indicated that accessions PER1010536, PER1010538, PER1010543, and PER1010547 were associated with high yield and early-maturity traits. Furthermore, multiple correspondence analysis and mixed data factor analysis demonstrated that the observed variability also depends on qualitative traits such as petal spot and bract color, supporting the concept of a complex genetic architecture. These findings provide a solid basis for the development of new cotton cultivars with improved productivity.