An　improved　Least　Squares　Reverse　Time　Migration　(LSRTM)　method　is　proposed　in　the　paper,　which　can　effectively　improve　convergence　speed　and　imaging　accuracy.　Firstly,　the　key　techniques　in　the　implementation　of　LSRTM　are　discussed.　Secondly,　a　condition　factor　is　introduced　in　the　iteration　process　of　conjugate　gradient　method.　Finally,　the　imaging　effect　and　performance　of　the　algorithm　are　analyzed.　The　experiment　results　indicate　that　it　can　speed　up　the　convergence　speed　and　improve　the　convergence　accuracy,　so　as　to　improve　the　imaging　effect.　Compared　with　the　conventional　LSRTM,　the　data　residual　of　improved　LSRTM　can　be　reduced　by　about　5%.　©　The　Authors,　published　by　EDP　Sciences　2021.