This paper describes an attempt to build a lexical database for the Yami language, an Austronesian endangered language. As the Yami language documentation and conservation projects have produced substantial corpora, we are now ready to construct the Yami online knowledge database based on the knowledge we have accumulated in the language. In this paper, we propose a model to build the WordNet-like Yami lexical semantics and database. The model is first described in detail, followed by an illustration of an ontology of fish in the implementation phase.