A field experiment comprising 45 sesame (Sesamum indicum L.) genotypes, including two checks (YLM-146 and TKG-22), was conducted during rabi 2025-26 at the dryland farm of Sri Venkateswara Agricultural College, Tirupati, Acharya N.G. Ranga Agricultural University, in a randomized block design with three replications. Genetic divergence among the genotypes was assessed for 14 seed yield and yield-attributing characters using Mahalanobis's D² statistic. Seed yield per plant (17.9%) contributed the maximum towards total divergence, followed by number of secondary branches per plant (17.1%), days to 50% flowering (13.2%) and 1000-seed weight (10.9%). Applying Tocher's method, the 45 genotypes were grouped into six non-overlapping clusters. Cluster I was the largest, comprising 28 genotypes, followed by cluster II with eight, cluster III with four and clusters IV and V with two genotypes each, while cluster VI, represented by the single genotype JCS-3993, indicated its distinct genetic constitution. The maximum inter-cluster distance was recorded between cluster III and cluster V (2803.64), followed by cluster V and cluster VI (2018.30) and cluster I and cluster V (1594.03), whereas the minimum inter-cluster distance occurred between cluster II and cluster IV (405.02). Cluster V recorded the highest mean values for number of capsules per plant and harvest index, while cluster II recorded the highest mean for seed yield per plant. As no single cluster combined all the desirable attributes, hybridization between genotypes drawn from the widely divergent clusters III, V, VI and I is suggested to generate superior transgressive segregants and to broaden the genetic base of cultivated sesame.