When　users　seek　for　Network　in　a　heterogeneous　cellular　network,　physical　cell　identification　(PCI)　is　used　to　distinguish　different　cellular　cells.　However,　the　number　of　PCIs　is　extremely　limited.　Moreover,　they　are　unable　to　satisfy　large-scale　random　deployment　of　small　cellular　cells　as　well　as　difficult　to　guarantee　users＇　quality　of　service.　Therefore,　this　paper　presents　a　fuzzy　hierarchical　clustering-physical　cell　identification　allocation　scheme.　The　proposal　is　based　on　the　＂activity＂　of　cellular　base　station　(CBS)　and　uses　the　Euclidean　distance　method　to　obtain　the　degree　of　similarities　between　the　CBSs.　After　dynamic　fuzzy　clustering,　the　optimal　clustering　results　are　obtained　by　using　the　analysis　of　variance.　In　this　paper,　high　priority　of　allocating　and　reusing　the　PCIs　are　given　to　those　CBSs　with　higher　activity.　Simulation　results　show　a　clear　increase　in　PCI　allocation　efficiency　and　reduction　in　PCI-conflict　and　PCI-confusion　possibilities　when　compared　with　the　existing　schemes.