With the continuous development of industrial automation, the application of industrial robots is rapidly increasing. Simultaneous Localization and Mapping (SLAM) for autonomous mobile robots in unknown environments have become a research hotspot. Currently, laser SLAM technology has made great progress, and many open-source laser SLAM methods have matured, but their performance in real-world environments is not confirmed. In this paper, three mainstream laser SLAM algorithms are applied to a tracked mobile robot system in an indoor environment to evaluate the algorithms' localisation performance in real environments. The experimental results show that the Gmapping algorithm exhibits better localization performancein small-scale scenarios, while the Karto and Cartographer algorithms have smaller differences and need to be selected according to the complexity of the real scene.