In exposing criminal acts, it is necessary to have a concrete piece of evidence, one of which is by using the role of fingerprints. Police fingerprint identification uses supporting tools such as the INAFIS Portable System (IPS). The process of reading fingerprints using this IPS tool, usually decreases image quality which causes obstacles in reading or analyzing fingerprint patterns. Nowadays digital image processing techniques have been widely used to conduct analysis and pattern recognition. This of course, can be used to perform fingerprint pattern recognition based on texture characteristics. In this study, a system was created that was able to recognize the types of fingerprint patterns as a form of digital technology development (image processing). In this analysis system research, the Gray Level Co-Occurence Matrix (GLCM) method is used by utilizing feature feature extraction by paying attention to the pixel relationship and Multi-Class Support Vector Machine (Multi-SVM) as an introduction to fingerprint pattern types. The data used in this study were fingerprint images of scanning results from IPS tools in the form of arch, left loop, plain whorl, right loop, and twinted loop patterns. Testing this analysis system produces varying degrees of success. The average success rate on this system uses training data which is 91.6% with the highest percentage success rate of 100% in the type of arch, left loop, and plain whorl pattern. While the average success using test data is 66% with the highest percentage success rate of 100% in the arch pattern type.