Cross-flow　filtration　mode　can　greatly　alleviate　the　membrane　fouling　due　to　the　presence　of　shear　stress.　In　this　paper,　a　new　flux　prediction　model　for　steady　cross-flow　microfiltration　has　been　established　by　considering　the　effect　of　shear　stress　as　function　of　the　geometry　of　filtration　module　and　the　suspension　properties　in　the　particle　mass　balance.　The　results　showed　that　this　model　possessed　higher　accuracy　in　both　laminar　and　turbulent　flow　regime　(relative　error　sigma　=　4.43%/1.91%)　than　Glen　Bolton＇s　model　(sigma　=　11.88%/6.42%)　and　Makardij＇s　model　(sigma　=　9.44%/12.92%).　Meanwhile,　the　validity　of　the　proposed　model　was　approved　by　using　other　suspensions　(yeast　suspension　and　activated　sludge　suspension)　(sigma　＜　3.43%)　and　other　membranes　(PVDF　and　PES　microfiltration　membrane)　(sigma　＜　4.63%).　Moreover,　new　criterion　relationships　between　Sherwood　number　(Sh),　Reynolds　number　(Re)　and　Schmidt　number　(Sc)　in　both　laminar　and　turbulent　flow　regime　were　also　established.