The invention discloses a Voronoi tree graph-based information retrieval visualization system and method. The method comprises the steps of extracting public attributes according to a given data set, constructing multiple trees by taking the attributes as root nodes, adding sub-nodes for the nodes with sub-attributes, adding auxiliary sub-nodes which are located in the gravity center of a Voronoi unit and represent themselves, and generating a Voronoi tree graph; displaying all layers of Voronoi regions of the Voronoi tree graph in sequence, selecting Boolean operators and attributes, generating new Voronoi regions for displaying Voronoi regions of the sub-nodes, continuously updating selection path trees, and generating a selection path tree which represents a Boolean operation expression; and traversing the selection path tree to obtain the Boolean expression, converting the Boolean expression into a corresponding database query language, performing a query in a database, and returning and displaying a data list meeting the condition. According to the method, same-level attribute multi-selection and cross-level attribute multi-selection are supported for querying data, so that the universality is enhanced.