Affinity　propagation　(AP)　is　a　clustering　algorithm　based　on　the　similarities　between　data　points　and　is　proved　to　be　a　fast　and　efficient　clustering　method　for　large-scale　data　sets.　But　for　some　data　sets　with　complex　cluster　structures,　it　cannot　produce　good　clustering　results.　In　this　paper,　a　novel　clustering　approach　based　on　the　combination　of　grey　relational　analysis　and　AP　algorithm　is　proposed.　The　similarities　between　data　points　are　described　by　the　balanced　closeness　degrees　of　their　attribute　sequences　to　improve　the　performance　of　AP　algorithm.　The　experimental　results　on　representative　data　sets　prove　the　superiority　of　the　new　approach　to　AP　algorithm　and　other　comparative　methods.