In this study, 15 morphological traits and 16 microsatellite markers were used to assess the morphological variability and structure of 68 (33 local and 35 foreign) almond accessions (Prunus dulcis (Mill.) D.A. Webb). Extensive phenotypic diversity was found among the accessions, and results indicated a high variation in leaf and fruit traits. Varieties were separated into two distinct groups with a similarity coefficient of 0.761. Morphological traits were categorized by principal component analysis into five components, which explained 86.5% of the total variation. Nut and kernel traits were dominant in the two first components, accounting for 49.4% of the variation. By contrast, leaf traits accounted for 18.4% of the variation in the third component. The results of molecular analysis (Bayesian clustering approach) did not correspond to morphological groupings, and the second approach was more discriminate. The combination of both approaches revealed the richness among the collected plant materials, which will be useful in breeding programmes of this species.