As a sub-type of ncRNA genes, micro RNA (miRNA) produces RNA that is implicated in the regulation of gene translation and expression. Human diseases can result from alternation in such non-coding RNA. Building a consolidated miRNA database repository thus becomes a crucial element for miRNA-related research. We have set up a dual objective in our work: To explore alternative approaches for miRNA data modeling, and to consolidate miRNA data in an integrated repository for future retrieval and analysis of miRNA data. We have explored three approaches for miRNA data modeling. For the same source data, we have implemented a relational database (converted from an initial entity-relationship diagram), an XML database, and an integrated XML-relational database. We have also compared their performance on the criteria of query performance and design flexibility. Though using entity-relationship modeling of the data provides access to the strength and stability of relational databases, our work concludes that XML information modeling and native XML databases provide geneticists the adaptability and flexibility required to model a complex, constantly changing domain such as miRNA research.