The　research　on　complex　networks　has　been　a　hotspot　in　many　fields,　among　which　community　structure　is　a　common　feature　in　complex　networks.　The　whole　network　is　composed　of　many　communities,　and　the　connections　between　nodes　in　the　same　community　are　very　tight,　while　the　connections　between　communities　are　relatively　sparse.　In　this　context,　this　paper　focuses　on　a　SWLPA　(Similarity　and　Weighted　Based　Label　Propagation　Algorithm)　label　propagation　algorithm　based　on　similarity　and　neighbor　node　weight.　This　paper　first　calculates　the　probability　of　label　propagation　according　to　the　similarity　between　nodes　and　the　weight　of　neighbor　nodes,　then　filters　labels　according　to　the　influence　of　each　node　and　carries　out　label　propagation,　and　finally　gets　the　result　of　community　division.　The　experimental　result　shows　that　the　method　effectively　improved　the　accuracy　of　community　division　results.