In breeding programs, it is necessary having knowledge of the relatedness and genetic diversity in germplasm pools. The spread of cultivated regions and the high levels of production indicates citrus importance in the global economy. Therefore, 110 citrus genotypes were evaluated using 12 ISSR markers. Overall, 154 polymorphic bands were scored with an average of 12.8 alleles per primer. The polymorphism percentage ranged from 57 for the ISSR1 to 82 for the ISSR9. Averages of polymorphic information content (PIC), marker index (MI), gene diversity index (Nei), Shannon index (I) and number of effective alleles (Ne) were 0.48 ± 0.002, 6.14 ± 1.17, 0.42 ± 0.11, 0.61 ± 0.12 and 1.78 ± 0.27, respectively. Based on genetic diversity statistics, the studied population had high genetic diversity, and four markers (ISSR11, ISSR9, ISSR4, and ISSR5) had more potential for differentiation of genotypes. Cluster analysis and model-based structure analysis, divided the genotypes into five groups and four subpopulations based on the Neighbor-Joining method (NJ) and Bayesian approach, respectively. Based on both analyses, grouping of unknown genotypes and control cultivars in the same group probably confirms the assumption of a common genetic background between these genotypes. Results from the two analyses showed that Pummelo (C. maxima), Mandarin (C. reticulate), and Citron (C. medica), as three true citrus species, separated in different groups. In addition to the three true species, at least one species or another genus of citrus relatives is involved in the genetic makeup of the studied population. In this study, although both used analyses were effective in completing each other's information, by considering the degree of genetic mixing and the information of the origin of the genotypes, the effectiveness of model-based structure analysis in evaluating genetic relationships could be achieved.