Genetic variability in germplasm collection is explored for its role in improving plant breeding strategies. In this study, we assessed the pattern of variation on 15 yield and bunch components in 2191 Elaeis guineensis dura germplasm from 13 populations (TZA01-TZA13) originating from Tanzania and planted in Malaysia in 1990. The data retrieved from the Malaysian Palm Oil Board-Breeding Information System (MPOB-BIS) were standardised prior to principal component analyses (PCA) and cluster analyses (CA) using SAS 9.4. First four principal component (PC) (PC1-PC4) having eigenvalue >1 accounted for 92.33% of the total variability with values 41.94%, 25.09%, 16.35% and 8.95%, respectively. PC1 has largest positive association with oil to dry mesocarp, oil to bunch and oil to wet mesocarp whereas largest negative association with kernel to fruit, kernel yield and kernel to bunch. PC2 is positively associated with mean fresh fruit bunch, mean average bunch weight, oil and kernel yield whereas negatively associated with shell to fruit and fruit to bunch. CA revealed high genetic variability lies within Tanzania germplasm collection. The combination of PCA and CA is a useful tool to select specific populations to establish core collection for better efficiency in oil palm improvement.