Forty four landraces of lowland rice were assessed for genetic diversity using D2 statistics based on various grain yielding parameters. Significant variability was found among the traits studied. Eight diverse clusters were made to set them into groups. The pattern of distribution of genotypes from different eco geographic regions into various clusters was at random indicating that geographical diversity and genetic diversity were not related. The maximum inter-cluster distance was observed between cluster III and IV (1033.15), followed by clusters II and VIII (935.07). Maximum intra-cluster distance was observed in cluster IV (731.01). Divergence analysis revealed significant variation among the parental lines for all the agro-morphological traits. Cluster analysis showed that superior cultivars like ‘Kanakchur’ from cluster IV; cv. ‘FR13A’, ‘Jaladhi2’, ‘Matla’ and ‘Purni’ from cluster III; cv. ‘Latisail’, ‘Pimpu Dibasa’, ‘Jeeringa Samba’ and ‘Shib Chato’ from cluster II; and cv. ‘ARC11210’, ‘Kumar Gore’, ‘NC678’, ‘Sada Mota’ and ‘Kantaranga’ from cluster VI may be taken in hybridization program, based on the genetic distance, relative contribution of characters towards total divergence and yield potential of genotypes. The characters viz., days to 50% flowering, plant height, panicle weight and number of grains panicle-1contributed maximum towards the divergence among genotypes taken for this study. So, allelic diversity among the genotypes and the concerned traits taken for this experiment should be taken into consideration while framing a rice breeding programme.