Barcode is a visual representation of information in form of bars and spaces. Barcode enable for description and identification of items accurately in a short time. The widespread use of barcode has significantly contributed for warehouses and retail product. Nowadays research about barcode is still an interesting topics especially from blurry, low contrast, low resolution and rotated barcode. This research is to investigate the possibilities of one-dimensional barcode recognition using Support Vector Machine (SVM) multiclass one-against-all with feature extraction using Principal Component Analysis (PCA) variation of principal component are 8, 12, 17, 25, 38, and 70 features. Datasets are taken from WWU Muenster Barcode Database University of Muenster Germany consists of 142 images from 13 types of barcode EAN-13. Before the SVM process, data are divided randomly into data training and data testing using cross validation repeated five times with ratio 2:1 so data training consists of 95 images and data testing consists of 47 images. Based on the best performance result, SVM is capable for classifying barcode accurately with accuracy (92.34±2.4) %. Based on computation time, the average of training time is about 3.21 seconds and testing time is about 0.66 seconds