ETL　tool　is　an　important　part　to　build　a　data　warehouse　and　data　centers,.　For　massive　data　processing,　this　paper　presents　an　intelligent　ETL　workflow　framework　based　on　the　distributed　computing　servers,　adding　an　intelligent　manipulative　module,　acquiring　the　data　of　the　system　efficiency　and　resources,　operating　data,　dynamically　adjusting　the　ETL　strategy,　and　doing　corresponding　data　segmentation　for　larger　jobs,　realizing　workflow　optimization　for　multi-machine　parallel　execution,　improving　operational　efficiency,　and　facilitating　error　recovery.　Intelligent　control　module　is　composed　of　the　monitor,　knowledge　base,　and　the　selector.　The　source　data　horizontal　partition　is　the　basis　and　difficulty　to　achieve　multi-machine　parallel.　©2010　IEEE.