Recently, in various fields XML has been become a standard for information exchange in internet. Therefore, the researches on data modeling for storing and fetching the XML documents have been made actively. However, existing researches have weak points that they can support neither versioning nor fast fetching of documents while changing documents in dynamic environments. In this paper, we propose four kinds of hybrid data modeling schemes that combine fragmentation model and nonfragmentation model. Our data modeling schemes are suitable to dynamic environments. We also present guidelines that our hybrid data modeling schemes can be used for various applications. We shaw through various experiments that our hybrid schemes partially outperforms the existing modeling schemes in terms of insertion time, storage space and retrieval time.