Sixty four accessions of groundnut (Arachis hypogaea L.) evaluated under late kharif season have been classified using Principal component analysis (PCA) based on correlation matrix yielding eigen values and eigen vectors. Fourteen principal components (PC) have been extracted using the mean performance of the genotypes, first ten principal components contributed over 95% of variation. Relative positive weights by each of the component to each single character has shown pod yield per plant and kernel yield per plant being given high positive weight by first principal component. Biplot of first two principal components showed characters viz., plant height, harvest index, pod yield, kernel yield per plant and oil yield per plant distinguishing among the accessions along the first principal component vector. Cluster analysis was performed based on first ten PC scores accounting more than 95% of variation which classified the sixty four accessions in to three clusters. Accessions in clusters 1 and 2 showed a wide range for several agronomic characters. This provides convenience in selecting superior accessions from each of these clusters for various yield contributing in the future breeding programs.