Microsatellite (SSR) markers originated from the expressed sequence tag (EST) region of genome are functionally more informative than the anonymous molecular markers for study of genetic diversity. With development of the EST-SSR markers in safflower, their utilization in genetic and breeding studies of safflower is now possible. The present study was aimed at estimating genetic variation in safflower using cultivars, lines, cultivated and wild populations with EST-SSR molecular markers. Intra population genetic variation and genetic purity of safflower genotypes were also evaluated. Forty eight safflower genotypes including 42 cultivated genotypes belong to Carthamus tinctorius and 6 wild accessions belong to C. oxyacanthus and C. lanatus species were used. The results indicated 42 of 109 EST-SSR primers pairs produced polymorphic bands with clear banding patterns. A total of 145 bands ranged from 100 bp to 500 bp were developed and scored. An average of 3.43 alleles per locus was detected. Dice similarity coefficients had a range of 0.26 to 0.98 indicating high genetic variation among genotypes. Dice similarity coefficients for the cultivated genotypes ranged from 0.68 to 0.98. High transferability of EST-SSR markers observed in present study makes them useful in cross-species genetic studies and basic evolutionary applications. The results of the present study showed that C. oxyacanthus accessions are more closely related to cultivated species than C. lanatus . Based on cluster and principal coordinate analyses, safflower genotypes were grouped into three main clusters C. lanatus , C. oxyacanthus and C. tinctorius . The 42 cultivated safflower genotypes divided into 4 subgroups. The subgroup one included 16 genotypes most of which obtained from Europe-America such as PI 537636, PI 258417, PI 198844 landraces and Hartman, Gila, Yinic cultivars. The subgroup two comprised 18 accessions originated from Palestine, Egypt and Mexico, as well as 15 Iranian cultivars, lines and landraces. Most Iranian accessions were belonged to this subgroup. The subgroup three contained 7 genotypes including a Syrian landrace and 6 Iranian genotypes (Sina cultivar and 5 Iranian landraces). PI 506426 landrace obtained from China separated solely into subgroup four. These results indicated that in most cases the cultivated safflowers divided into the subgroups based on their country of origin and genetic background. However, there was slight inconsistency between country of origin and subgroup of some genotypes. Results of analysis of molecular variance showed a significant difference between Iranian and exotic genotypes. Although, the fixation index value F ST = 0.05 observed in the present study suggests a relatively low level of genetic differentiation between the two groups. The results of population genetic structure of four safflower population with four SSR primers indicated that Isfahan-W and C4110 populations had the highest and lowest values of observed and expected heterozygosity (gene diversity), respectively. The fixation index (F ST ) values, indicating the differentiation between subpopulations, ranged from 0.21 to .052 with an average of 0.39, which shows existence of high differentiation between used populations. A high value of gene diversity was observed in Arak landrace, indicating the potential of landraces as origin of selection to produce cultivars and breeding lines. In conclusion, these markers were effective for evaluation of genetic variation and population genetic analysis in the safflower species. Key words: Safflower, Genetic variation, Express sequence tag (EST), Simple sequence repeat (SSR), Carthamus