An investigation on divergence, variability and associations among yield related traits were studied using sixty four traditional rice cultivars. Based on D2 statistics, all the 64 cultivars were grouped into fourteen clusters. Of which cluster I (44) had more number of cultivars fallowed by cluster III (8) and remaining were solitary. Maximum inter cluster D2 distance was observed between cluster X and XIV (2056.50) inferring, crosses between these two clusters could exploit maximum heterosis. Whereas, maximum intra cluster distance was observed in cluster III (225.63) indicating, hybridization involving genotypes within the same cluster may result in good cross combinations. Days to maturity contributed maximum (36.41%) towards divergence, followed by straw yield per plant (19.54%). Genetic variability studies on yield and yield related parameters had confirmed the existence of high genetic variation and possibility of improvement of yield related traits based on genetic parameters like Genotypic Coefficient of Variation, Heritability and Genetic Advance. Genotypes such as Sanna mundaga, Kaase bai and Champakali which are genetically variable and high yielders over local check varieties could be utilized in crop improvement programme for enhancement both qualitative and quantitative traits. Plant height (0.1977*) and test weight (0.2559*) showed positive and significant correlation with grain yield at phenotypic level. Phenotypic path coefficient analysis had revealed the highest positive and direct effect of days to maturity (0.5107) followed by harvest index (0.3110) on grain yield. Hence, selection based on these traits could help to bring simultaneous improvement of yield and its components.