To　deal　with　the　prediction　of　seismic　landslide　hazard,　based　on　information　fusion,　a　methodology　was　proposed　for　predicting　seismic　landslide　using　the　evidence　theory　which　can　reflect　the　comprehensive　influences　of　different　factors.　Six　indices　related　to　the　occurrence　condition　of　seismic　landslide　were　taken　into　account　as　evidence　in　the　proposed　method,　including　the　coefficient　of　rock　weathering,　earthquake　intensity,　fault　density,　drainage　density,　relative　height　and　the　mountain　slope.　The　basic　probability　was　objectively　constructed　using　entropy　weight　grey　incidence.　The　seismic　landslide　prediction　model　was　built　by　evidence　theory　and　entropy　weight　grey　incidence.　The　results　show　that　the　method　has　relatively　high　accuracy.　Because　of　its　reliability　and　rationality,　this　method　can　satisfy　the　planning　on　earthquake　resistance　and　hazardous　prevention　generally.　©　2016,　Central　South　University　of　Technology.　All　right　reserved.