Accurate　flux　prediction　in　constant　pressure　cross-flow　microfiltration　is　of　great　significance　for　industrial　cleaning　process.　In　this　study,　a　new　combined　mechanism　(cake　filtration　and　complete　blocking)　model　was　developed　to　accurately　predict　the　flux　decline　in　constant　pressure　cross-flow　microfiltration,　which　based　on　the　Bolton＇s　model　with　considering　both　the　cake　resistance　and　the　available　membrane　frontal　area　tending　to　be　a　constant　after　a　certain　period　of　filtration　by　using　both　cake　filtration　equilibrium　coefficient　(K-d)　and　the　steady　frontal　membrane　area　(K).　The　proposed　model　was　validated　by　different　membranes　and　various　feed　suspensions　for　both　laminar　and　turbulent　regions.　The　results　showed　that　the　new　cake-complete　model　possessed　good　prediction　accuracy　for　different　membranes　(0.1　mu　m　PAN,　PVDF　and　PES　membranes)　(R-2　＞=　0.9603)　and　different　feed　suspensions　(yeast,　kaolin　and　activated　sludge　suspensions)　(R-2　＞=　0.9309).　Meanwhile,　the　prediction　accuracy　of　the　proposed　model　was　higher　than　that　of　others　and　the　sequence　of　average　adjust　R　square　((R-2)　over　bar)　was:　0.9696　(the　proposed　model)　＞　0.8755　(Silva＇s　model)　＞　0.8193　(Bolton＇s　model)　＞　0.6141　(cake　model).　In　addition,　low　Reynolds　number,　high　concentration　and　big　pressure　would　decrease　the　available　membrane　frontal　area　and　increase　K-d,　which　ultimately　leaded　to　more　serious　membrane　fouling.