Abstract: Exotic cultivars introduced from abroad and local selections from seedling population of walnut were subjected to Non-hierarchical Euclidean cluster analysis based on 12 nut and kernel characters. Studied accessions were grouped into four clusters, where cluster 1 and 3 had majority of the accessions. The clustering pattern of walnut genotypes belonging to the same eco-geographical region revealed their distribution in more than one cluster showing between geographic and genetic diversity. The first component presented maximum eigen root value and per cent variation. The mean intra and inter cluster distance (D) revealed that cluster 4 had highest intra cluster distance (2.206), while the inter cluster distance was maximum between cluster 3 and 4 (5.806). Maximum mean nut yield, nut length, nut weight and kernel weight was recorded in cluster 4, whereas, maximum kernel percentage and fat percentage was recorded in cluster 1 and 2, respectively. Minimum shell thickness was observed in cluster 2. The accessions of cluster 3 and 4 were found highly diverse from each other and will give better segregants after hybridization which can be used as a parent in further breeding programme.