This research was carried out to determine the morphological diversity of barley accessions in Ethiopia via an alpha lattice design with two replications. All the qualitative data were subjected to statistical analysis via Minitab 17^th Edition and SPSS software. Principal component (PC) analysis, cluster analysis, the Shannon–Weaver diversity index (H'), and distribution frequency were used for the qualitative traits. The frequency distribution of the phenotypic traits revealed a considerable degree of genetic variability among the accessions. The estimate of H' ranged from 0.83 (glume hairiness) to 0.99 (lemma awn). The highest value of H’ in the zones of origin was (H’=1.00) for the kernel row number in Shewa, growth habit and spike attitude in Gojam, glume hairiness in Harerge, and lemma awn in Semen omo. For the altitude classes, the trait lemma awn had the highest value of H’ (H’=1.00) in accessions collected at an altitude of ≥2600 m.a.s.l. Among the total variation among the populations, 71% was found within the zones of origin, and 75% was found within the altitude classes. The ANOVA for H’ values on the basis of the altitude class and zones of origin showed very highly significant (p≤0.001) variation to significant (p≤0.05) variation. On the basis of eigenvalues greater than 1, the first 3 PCs had cumulative contribution rates of 70.6%, 77.4%, and 95.7% of the total variation among accessions, zones of origin, and altitude classes, respectively. Cluster analysis revealed four distinct groups among the zones of origin and altitudinal classes. In general, traits such as kernel color, lemma color, and growth habit in the zone of origin and traits such as lemma color, lemma awn, and lemma color at physiological maturity in altitude classes were the most determinant traits among the barley accessions and could be used as selection criteria for crop improvement.