Plagiarism is the practice of taking someone else’s work or ideas and passing them off as one’s own. In this paper, we show automatic plagiarism detection using two methods. First method is MD5 message-digest algorithm and second method is words comparison. Our experiments showed that first method succeeded in detecting same paragraphs (or with a slight modification) with 100% accuracy. This kind of plagiarism occurs when one copies and pastes other’s paragraph into his/her own writing. Our second method succeeded in detecting in average of 40.7% similarity between two paragraphs which one is paraphrased version of the other. We propose that these two methods combined could make a novel classifier for plagiarism detection. [pdf]