Ontologies are an essential resource to enhance the performance of an information processing system as they enable the re-use and sharing of knowledge in a formal, homogeneous and unambigous way. We propose here a method for the automatic learning of an taxonomic ontology based on thai text corpora. To build the ontology we extact terms and relations. For the former we use a shallow parser, while for the extraction of taxonomic relations we use item lists, i.e. bullet list contain an ontological term and to solve the problems of embedding of lists and the boundaries of the list. Then, we extract the hypernym term of the item list by using the lexicon and the contextual features of the candidate term of the hypernym. The accuracy of the otonology extraction from item list is 0.72.