The　technology　of　Multi-view　images　has　been　found　wide　application　in　3D　reconstruction,　surveillance　systems　and　MRI.　However,　the　volume　of　multi-view　data　is　fairly　large　and　sometimes　the　time　is　also　limited.　In　order　to　cut　the　data　and　keep　good　image　quality,　we　propose　a　multiple-image　pattern　low-rank　tensor　algorithm　based　on　compressed　sensing,　in　which　we　propose　to　exploit　the　model　of　tensor　to　find　the　property　of　low　rank　and　extend　it　into　multiple-image　pattern.　In　our　work,　we　further　propose　an　efficient　algorithm　to　solve　the　low-rank　tensor　method　using　the　truncated　HOSVD　method　and　alternative　direction　multiplier　method　technique.　Experimental　results　demonstrate　that　our　method　outperforms　previous　methods　with　almost　1DB　higher　than　the　PSNR　of　existing　results.