Identifying　influential　nodes　is　of　theoretical　significance　in　network　immunization　which　is　one　of　important　methods　to　prevent　virus　propagation　through　protecting　the　influential　nodes　in　a　network.　Lots　of　methods　have　been　proposed　to　find　these　influential　nodes　based　on　the　topological　characteristics　of　a　network　(e.g.,　degree,　betweenness　or　K-shell).　Whereas　due　to　the　diversity　of　network　topologies,　these　methods　are　not　always　effective　in　identifying　influential　nodes　in　any　benchmark　networks.　We　combine　the　advantages　of　existing　methods　based　on　attribute　ranking　and　propose　a　universal　ranking　method,　namely　MAF　(Multiple　Attribute　Fusion),　to　identify　influential　nodes　from　a　complex　network.　We　compare　the　efficiency　of　our　proposed　method　with　existing　immunization　strategies　in　different　types　of　networks.　Simulation　results　in　the　interactive　email　model　show　that　the　immunized　nodes　selected　by　MAF　can　restrain　virus　propagation　effectively.