Ontology bootstrapping is based on a set of textual sources, such as web services, must address the problem of multiple and largely isolated concepts. Web information extraction is essential to many application areas, such as ontology mapping and bootstrapping semantic Web, etc. This paper shows that instances are usually backed by fundamental relational databases in majority Web sites. Based on this examination, two approaches for Instance Extraction are illustrated in this paper. This paper is to bootstrap instances, rich ontologies from Web documents.