Introduction. One of the most important tasks of forensic ballistics examination is the identification of weapons from marks of the striker. This problem is usually associated with the searching for matching traces from cartridge case repository. Ballistic identification systems (such as POISC, TAIS, Condor, Arsenal and others) are used to help expert in such task. All of them use automatic algorithms for searching matching traces based on the calculation of the cross- correlation function. In this paper cluster analysis methods is used to form a priority list of striker traces images which contains individual attributes in the form of large irregularly shaped areas. Theoretical part. In the theoretical part the algorithm is described for extraction and compassion of fire pin marks. Then it is used for forming a priority list. Experimental part. The developed algorithm was applied to cartridge case repository consisting more than 100 images of the strikers traces of 24 weapon samples. Conclusion. The algorithm allows to form a priority list of 20 images, which includes paired tracks (if bullets repository contains it) with a probability of 100%. Thus it helps to reduce amount of images for comparison by 80%. The proposed method can significantly reduce the search time. To f urther reduce the priority list should be used more intensive methods, such as contour analysis or singular points comparing.