Early　warning　is　an　effective　way　to　eliminate　or　reduce　the　losses　and　negative　impacts　caused　by　hazards　in　advance,　which　has　important　practical　significance　in　real　world　situations.　It　has　drawn　great　attention　from　government　and　academy　in　recent　years.　Information　related　to　hazards　plays　an　important　role　in　early　warning　process,　there　are　different　types　of　such　information　in　real　world　situations.　However,　existing　early　warning　studies　employ　just　single　information　type　to　describe　the　hazards,　it　might　not　describe　the　hazards　properly　and　comprehensively　due　to　the　complex　and　uncertain　of　real　world　situations.　Therefore,　it　seems　necessary　and　reasonable　to　consider　different　types　of　information　related　to　hazards　in　the　early　warning　process.　To　fill　up　the　gap,　this　study　proposes　a　novel　early　warning　method　that　considers　heterogeneous　information　from　both　quantitative　and　qualitative　contexts　including　experts＇　hesitancy.　On　such　basis,　fuzzy　evidential　reasoning　approach　is　utilized　due　to　its　capacity　not　only　to　fuse　different　types　of　information,　but　also　to　keep　as　much　information　as　possible　during　the　fusion　process.　Finally,　an　illustrate　example　together　with　related　discussions　are　presented　to　demonstrate　the　validity,　feasibility　and　novelty　of　the　proposed　method.　The　proposed　early　warning　method　can　provide　more　safety　status　information　of　early　warning　objectives,　which　can　be　used　to　not　only　forecast　and　warn　different　types　of　hazards,　but　also　conduct　the　risk　analysis　and　assessment.　It　has　widely　potential　application　values　in　real　world　situations.