Sentiment　analysis　is　a　basic　task　of　natural　language　processing,　while　aspect　level　sentiment　analysis　is　an　important　topic　in　sentiment　analysis.　In　the　same　sentence,　different　words　have　different　influence　on　the　sentiment　polarity　of　aspect,　so　the　key　to　solve　the　problem　is　how　to　build　a　relation　model　between　the　aspect　and　the　words　in　the　sentence.　In　this　paper,　by　using　two　recurrent　networks,　we　built　a　model　for　sentence　and　introduced　attention　mechanism　to　fuse　aspect　information,　so　as　to　achieve　a　better　effect.　An　experiment　on　public　dataset　show　that　the　proposed　algorithm　obtain　a　better　result　without　carrying　out　complex　feature　engineering.　©　2019　Published　under　licence　by　IOP　Publishing　Ltd.