A　new　modified　Min　Sum　decoding　algorithm　based　on　parameter　estimation　theory　for　Low　Density　Parity　Check　(LDPC)　codes　is　presented　in　this　paper.　First,　Linear　Minimum　Mean　Square　Error　(LMMSE)　estimate　criterion　is　used　to　model　the　magnitude　of　check-to-variable　message　and　calculate　estimated　parameters,　furthermore　the　golden　section　search　algorithm　is　used　to　speed　up　getting　the　boundary,　with　which　the　parameters　can　be　determined.　Then,　the　parameters　are　modified　further　on　the　basis　of　the　relationship　between　the　iteration　number　and　the　estimated　parameters.　Finally,　the　same　estimated　parameters　are　used　for　different　signal　to　noise　ratio　(SNR)　to　ensure　the　decoding　performance　and　reduce　the　hardware　complexity.　The　simulation　results　show　that　a　gain　of　0.3-0.5　dB　can　be　achieved　in　comparison　with　other　kinds　of　modified　Min　Sum　algorithms,　which　is　very　close　to　that　of　BP　algorithm.　What＇s　more,　the　algorithm　has　the　following　characteristics:　low　computation　complexity,　small　decoding　delay,　easiness　of　calculation　for　estimated　parameters,　and　simplicity　in　hardware　implementation.　©　2014　Elsevier　GmbH.