Today, mathematical modeling is an important tool in scientific research. In mathematics, publications are containers for mathematical research data containing models and mathematical objects, theories, theorems, proofs, methods, algorithms, and applications. However, it is a nontrivial task to extract detailed information about models and other research data from publications. Especially, the information about mathematical models is given in different form, e.g., by names or in “natural mathematical language” (a combination of text and mathematical formulae), and is often incomplete. Moreover, different communities use different terminologies and notations for the same models. These deficits make it difficult to find and reuse existing models and lead to repeated reinventing the wheel. This results in an project idea to analyze and collect existing information about mathematical models in publications and to develop concepts for a semantic presentation of them. In the talk, the presentation of research data in mathematical publications will be analyzed and first ideas for a special model database are suggested.