The PCR-SSR technique was used to detect nuclear DNA diversity in five wild populations of Prunus avium from deciduous forests in Italy, Slovenia, and Croatia and 87 sweet cherry accessions from different geographical areas that have been maintained in the sweet cherry collection in Italy. This sweet cherry collection includes local accessions from the Campania Region as well as accessions from different countries. Twenty-eight microsatellites, previously developed in this species, generated polymorphic amplification products. Between 2 and 14 alleles were revealed for the polymorphic loci studied, with the expected heterozygosity ranging from 0.045 to 0.831. The total probability of identity was 56.94 x 10-18. A model-based Bayesian clustering analysis identified nine distinct gene pools in cultivated P. avium. The probability that wild populations were assigned to cultivated gene pools indicated that three gene pools accounted for the genomic origin of 53% of P. avium sampled. A dendrogram was generated using UPGMA (unweighted pair group method with arithmetic averages) based on Nei genetic distance analysis. This dendrogram classified most of the genotypes into one major group with an additional group of five accessions. The results indicate that this set of SSRs is highly informative, and they are discussed in terms of the implications for sweet cherry characterization.