The total of 27 genotypes/selections belonging to Juglans genus were assessed, to estimate the extent of genetic relationship by morphological and molecular characterization using UPOV descriptors (International union for protection of new varieties of plants) (UPOV, 1999) and RAPD (Random Amplified Polymorphic DNA) markers. Multivariate analyses which include Principal Component analysis and Cluster Analyses, was employed to examine the extent of phenotypic variation in the genotypes/selections of walnut. Based on morphological data, the genotypes were grouped into nine clusters. Principal component analysis was performed to assess relationship among different components. First two principal components accounted for 98% of total variation. The total of 16 primers were used in this study out of which 14 yielded scorable amplification patterns which resulted in the generation of DNA fragments that ranged from 500-1500 base pairs. RAPD data was used to calculate similarity matrix using Jaccards coefficient and similarity coefficient ranged from 0.05 (SKAUW-15 & SKAUW-7) to 0.86(SKAUW-008 & SKAUW-040). Dendrograms revealed that all the walnut genotypes/selections were grouped into seven major clusters Polymorphic Information Content (PIC) value ranged from 0.11 (A-15)-0.39 (A-11) and resolving power ranged from 0.12-1.29. Cluster analysis based on molecular data did not show any correlation with morphological characters based on Mantels test (r=0.09, P=0.87).