Experimental　evidence　indicates　that　a　majority　of　paths　generated　by　static　analysis　tools　are　found　to　be　infeasible.　As　structural　testing　is　an　integral　part　of　many　software　testing　activities,　detecting　infeasible　paths　at　an　early　phase　can　greatly　improve　the　efficiency　of　many　structural　testing　techniques.　Since　existing　approaches　commonly　handle　infeasible　paths　from　the　function　units,　a　novel　approach　is　proposed　in　this　paper　to　detect　infeasible　paths　interprocedurally.　We　use　a　map　data　structure,　called　interprocedural　symbolic-propagation　mapping,　to　model　the　value-passing　process　at　each　call　site　along　the　generated　interprocedural　paths,　and　then　a　hybrid　method　is　utilized　to　determine　the　feasibility　of　each　given　path　on　the　fly,　which　combines　interprocedural　dataflow　analysis　with　the　symbolic　propagation　technique　without　applying　constraint　solvers.　Experimental　results　prove　the　effectiveness　of　the　proposed　approach.