Abstract: To support decision-making in enterprise,a cube data reconstruction model is presented to put the heterogeneous data from different sources in order.At first,all kinds of data for parts and sub-assemblies are concatenated into the form of individual matrices,and then row-matching and column-merging is used to combine the matrices into a synthetic one,showing the comprehensive information of design,process,production and management.However,the synthetic matrix is still huge.Projection transformation and selection transformation are applied to pick out data items necessary for decision making towards the goal.Finally,the overall cube data model for decision support is formulated by merging the synthetic matrix of every part and subassembly.