An　ultrasonic　reverse　time　migration　imaging　method,　based　on　high-order　singular　value　decomposition,　is　proposed　in　the　study　to　solve　the　problems　of　low　signal-to-noise　ratio　(SNR)　and　excessive　artifacts　in　defect　ultrasonic　detection　imaging　results　of　materials　with　high　noise　levels.　In　this　method,　based　on　the　3D　structural　properties　of　the　ultrasonic　full-matrix　capture　data,　higher-order　singular　value　decomposition　is　directly　performed　with　the　3D　data.　The　method　overcomes　the　difficulty　in　selecting　the　number　of　singular　values　in　the　original　singular　value　decomposition　noise-reduction　algorithm　and　realizes　the　one-step　noise　reduction　processing　of　all　the　signals.　Subsequently,　the　reverse　time　migration　imaging　is　performed　in　the　frequency　domain,　and　high-quality　acoustic　images　are　obtained.　The　effects　of　the　number　of　array　elements,　the　center　frequency　of　the　excitation　signal,　and　the　number　of　defects　on　the　denoising　effect　of　the　algorithm　are　investigated.　It　was　experimentally　demonstrated　that　the　method　could　suppress　the　interference　of　noise　signals　and　significantly　improve　the　imaging　SNR　compared　with　total　focusing　method　and　the　reverse　time　migration.