This study was conducted to investigate the genetic diversity of cultivated safflower germplasm and to identify the relationships among wild safflower species using molecular marker and morphological characters. For this reason inter simple sequence repeat (ISSR) was used to study 107 accessions from 7 species of Carthamus including 79 genotypes of cultivated species ( C. tincterious ), 9 genotypes of C. lanatus, 7 genotypes of C. oxyacanthus , 4 genotypes of C. glaucus and one genotype from each of the species C. boissieri, C. dentatus and C. palaestinus were used. Result indicated that a total of 508 markers were amplified and the rate of polymorphism was 72 percent (368 markers) The size of the bands ranged from 300 bp to 2000 bp and the number of amplified products ranged from 14 to 25 per primer. For data analysis, ISSR bands throughout the gel profile were scored as present (1), absent (0) or ambiguous (9) at least twice. The NTSYSpc software was used to generate genetic similarity (GS) matrixes, create dendrogram and corresponding cophenetic matrixes, and calculate cophenetic correlation. Result indicated that the GS values ranged from 0.19 to 0.98 among total of the genotype and ranged from 0.29 to 0.98 for genotypes of C. tinctorius . Cluster analysis showed that C. palaestinus were clearly clustered with C. tinctorius in a single group. Similarity coefficient using Nei method confirmed that C. palaestinus had more genomic similarity (90%) with C. tinctorius than other studied species especially C. oxyacanthus as one the wild relatives. This results support the theory of the origin of safflower by wild species, C. palestinus . Result from cluster analysis and principal component analysis was able to put each of the 7 species of safflower in separate cluster with the exception of C. palaestinus which was clustered near the cultivated genotypes. Results of principal component analysis showed that the 2 first components justified about 37 percent of total variation indicating the primer loci may have low genetic linkage with each others. For a better illustration of genetic similarities between cultivated accessions, cluster analysis was performed of these accessions. All of the cultivated accessions were clustered into 5 groups. The first group included genotypes from Europe, Asia, Africa and America. Accessions of Egypt and Sudan and two accessions from Australia formed the second group. The third group included 27 genotypes of the regions of central Asia, southwest Asia and America. An accession of C. palaestinus was also clustered in this group besides the accessions of Jordan, America and India. The fourth group was composed of genotypes from the Middle East and central Asia indicating that safflower has been introduced to areas of west Asia. Fifth groups, including the samples from the far east and Uzbekistan. Results showed that there was considerable congruity between ISSR grouping and geographical distribution of accessions. There were significant differences between genotypes of C. tinctorius introduced from the six diversity centers. According to the gene diversity information index of Nei and Shannon, high variation was found between samples belonging to Middle East which strengthens the hypothesis that this region is the origin of safflower. On the other hand, center of the far east showed the least variation. Cluster analysis based on morphological data performed using the NTSYS pc software has almost been able to separate different species from each other. However low correlation was found between morphological and ISSR data. In addition, the result indicated high similarity between C. palaestinus and C. tinctorius indicating possibility of breeding the cultivated species using interspecies hybridization. Key word: genetic diversity, phylogenetic relationship, center of similarity, ISSR marker