In the last couple of years, unsupervised machine learning has emerged as a strong tool for the analysis of complex patterns and relationships in data. Its application in the assessment of anthropometric characteristics and biological maturity is of great value for gaining insight into athlete development and performance. The present study intended to characterize the body composition, morphology and maturity status of young athletes; classify them according to anthropometric measurements; and predict their adult stature. The sample included 138 male athletes of varying regional settings in Bangladesh, all of whom were aged between 12 and 14 years. The anthropometric measurements used are those prescribed by ISAK. Key variables involved were body mass, stature, girths, and skinfold thickness. PAM clustering was used to identify patterns; morphological variable associations with maturity status were determined using statistical methods. Results identified two clear clusters, where Cluster 1 was dominated by late-maturing athletes and Cluster 2 was comprised of averagely matured athletes. Cluster 2 athletes were more physically robust, higher in body mass, and more advanced in maturity, hence having greater athletic potential. These findings constitute an effective contribution to the identification of talented athletes and specific intervention strategies for supporting athletic development in Bangladesh.