The　state　of　high　speed　railway　infrastructure　affects　the　safety　of　train　operation.　Sensor　network　technology　makes　it　possible　to　realize　real　time　and　all-weather　detection　for　high　speed　railway　infrastructure.　According　to　the　application　requirements　of　the　sensor　network　for　detecting　high　speed　railway　infrastructure,　a　three-layer　network　communication　model　is　proposed,　which　is　composed　of　End　Device　(ED),　Repeater　Point　(RP)　and　Access　Point　(AP).　Given　the　types　and　amount　of　EDs　in　a　detection　area,　an　integer　programming　model　is　built　to　maximize　the　bandwidth　utilization,　and　the　physical　topology　optimization　problem　of　sensor　network　is　converted　into　multiple　integer　programming　one.　For　two　cases　of　with　no　sensor　network　and　with　some　sensor　networks　in　a　detection　area,　a　PSO　algorithm　based　solving　scheme　is　proposed.　An　application　example　with　60　EDs　is　shown　to　validate　the　proposed　method.　The　optimal　numbers　of　AP　and　RP　and　their　corresponding　EDs　combined　strategies　are　all　obtained　within　50　iterated　operations.　Numerical　analysis　results　show　that　the　proposed　communication　model,　physical　topology　optimization　method　and　algorithm　can　solve　the　physical　topology　optimization　problem　of　high　speed　railway　infrastructure　detection.　In　addition,　they　can　effectively　lower　the　cost　of　sensor　network　and　determine　the　optimal　links　for　network　communication.　©,　2015,　Chinese　Academy　of　Railway　Sciences.　All　right　reserved.