In　this　paper,　the　PCA-PSOBP　neural　network　has　been　put　forward　to　model　ultrafiltration　of　printing　and　dyeing　wastewater.　Firstly,　Principal　Component　Analysis　(PCA)　was　applied　to　reduce　the　dimensions　and　correlations　of　input　parameters.　Secondly,　the　PSOBP　was　used　to　optimize　the　weights　and　thresholds　of　the　neural　networks,　in　which　weights　of　BP　neural　network　were　adjusted　by　particle　swarm　optimization　(PSO)　rather　than　traditional　gradient　descent　method.　Based　on　experimental　data,　simulations　are　performed　with　MATLAB.　The　results　showed　that　PCA-PSOBP　neural　network　has　a　faster　convergence　speed　and　a　better　agreement　with　the　real　data　than　traditional　BP　neural　network.　©　2010　IEEE.