Fire　harms　human　life　and　properties.　Fire　detection　technology　helps　people　prevent　fire　and　reduce　losses.　Traditional　physical　sensing　fire　detectors　have　the　disadvantages　of　untimely　alarms　and　high　false　alarm　rates.　With　the　popularization　of　video　surveillance,　the　video　detection　of　fire　has　been　studied.　This　article　reviews　the　development　process　of　video　detection　fire　technology　in　recent　years,　and　the　two　main　directions　of　traditional　algorithms　and　deep　learning　are　analyzed.　The　characteristics　are　compared　to　understand　the　current　shortcomings　and　limitations.　©　2022　IEEE.