Plagiarism is the act of taking essay or work of others, and recognize it as his own work. Plagiarism of the text is very common and difficult to avoid. Therefore, many created a system that can assist in plagiarism detection text document. To make the detection of plagiarism of text documents at its core is to perform string matching. This makes the emergence of the idea to build an algorithm that will be implemented in RTG24 Comparison file.txt applications. Document to be compared must be a file. Txt or plaintext, and every word contained in the document must be in the dictionary of Indonesian. RTG24 algorithm works by determining the number of same or similar words in any text between the two documents. In the process RTG24 algorithm has several stages: parsing, filtering, stemming and comparison. Parsing stage is the stage where every sentence in the document will be broken down into basic words, filtering step is cleaning the particles are not important. The next stage, stemming is the stage where every word searchable basic word or root word, this is done to simplify and facilitate comparison between the two documents. Right after through the process of parsing, filtering, and stemming, then the document should be inserted into the array for the comparison or the comparison between the two documents. So it can be determined the percentage of similarity between the two documents.