Because　nonlinear　dynamic　systems　are　various,　they　are　still　not　well　solved　via　the　traditional　identification　methods.　This　paper　supposes　that　an　original　dynamic　system　is　described　by　Hammerstein　model　initially.　Its　transfer　function　can　then　be　changed　into　a　simple　form　via　expansion,　thus　generating　an　intermediate　model　and　its　parameters　are　obtained　by　an　impoved　particle　swarm　optimization.　Finally,　the　parameters　are　gotten　by　the　corresponding　parameters＇　relationships.　Accordingly,　the　original　system　is　identified.　To　demonstrate　the　feasibility　of　the　proposed　method,　illustrative　examples　are　included.