In the present investigation, 20 genotypes of Tunisian castor were analyzed using 5 start codon targeted (SCoT) markers. These primers produced total 33 fragments across 20 Tunisian castor genotypes, of which 26 (78.70 %) were polymorphic with an average of 5.20 polymorphic fragments per primer and number of amplified fragments ranged from 4 (SCoT 66) to 9 (SCoT 65). The polymorphic information content (PIC) value ranged from 0.652 (ScoT 65) to 0.816 (SCoT 61) with an average of 0.738. The dendrogram based on hierarchical cluster analysis using UPGMA algorithm was prepared. The hierarchical cluster analysis showed that the Tunisian castor genotypes were divided into 3 main clusters. Cluster 1 contained 4 castor genotypes from region Gabes (G-1, G-2, G-4 and G-5). Cluster 2 contained 4 castor genotypes from region Mornag (M-1, M-2, M-4, M-5). Cluster 3 contained 12 Tunisian castor genotypes and was divided into subcluster 3a, 3b, and 3c. Subcluster 3a contained one genotype from region Mornag (M-3), one genotype from region Mednine (MD-5) and 2 genotypes from the region Kebili (K-2 and K-3). Subcluster 3b contained 2 castor genotypes from region Kebili (K-1 and K-5). Subclaster 3c contained five genotypes from region Ksar jedid (KJ-1, KJ-2, KJ-3, KJ-4, KJ-5) and one genotype from region Kebili (K-4). Two Tunisian castor genotypes of 3c subcluster (K-1 nad K-4) were genetically the closest. We can assume that they have a close genetic background. The present study shows the effectiveness of employing SCoT markers in analysis of castor and would be useful for further studies in population genetics, conservation genetics, and genotypes improvement.