Methodology of ontology building based on Web resources will not only reduce significantly the ontology construction period, but also enhance the quality of the ontology. Remarkable progress has been achieved in this regard, but they encounter similar difficulties, such as the Web data extraction and knowledge acquisition. This paper researches on the characteristics of ontology construction data, including dynamics, largeness, variation and openness and other features, and the fundamental issue of ontology construction - formalized representation method. Then, the key technologies used in and the difficulties with ontology construction are summarized. A software Model-OntoMaker (Ontology Maker) is designed. The model is innovative in two regards: (1) the improvement of generality: the meta learning machine will dynamically pick appropriate ontology learning methodologies for data of different domains, thus optimizing the results; (2) the merged processing of (semi-) structural and non-structural data. In addition, as known to all wetland researchers, information sharing is vital to wetland exploitation and protection, while wetland ontology construction is the basic task for information sharing. OntoMaker constructs the wetland ontologies, and the model in this work can also be referred to other environmental domains.