Abstract: Multi-sensor systems have extensive applications in various fields such as air defense and anti-missile systems, unmanned collaboration, intelligent transportation, and multi-target tracking. The multi-sensor collaborative localization and tracking methods based on Random Finite Set theory differ from traditional data association algorithms in that they can better cope with complex environments such as time-varying targets and dense clutter. This paper systematically reviews the current main research ideas in multi-sensor collaborative localization and tracking. It addresses practical issues faced in multi-sensor fusion, including communication burden, unconventional parameter processing, and inconsistent sensor fields of view, and summarizes corresponding solution strategies. Furthermore, it analyzes the applicable conditions, advantages, and disadvantages of typical algorithms. Based on the current challenges and issues, the paper also provides an outlook on future research directions.