Aim of the research is to model the system for cataloguing in the UNIMARC format which does not require specific knowledge of cataloguing formats. The Unified Modeling Language (UML 2.0) is used for the specification of both the information requirements and the architecture model of system for cataloguing. The research finding is a model of system for cataloguing in UNIMARC format by which authors themselves can process their bibliographic entries without knowing UNIMARC format. Bibliographic records formed in this way can be further processed according to adopted standards by librarians. Limitation of research results lies in the additional modeling of graphic user interface for the purpose of changing input data set of library documents. This limitation can be surpassed by specifying the input data set which would be used for automatic generation of appropriate user interface. The specification could be realized by means of XML Schema language. Practical usage of the research findings is the basic for the implementation of a Web application intended for the creation of electronic catalogues and bibliographies of researchers and scientific institutions. Integration of the catalogue into BISIS makes it publicly available through a standard user interface for searching bibliographic records on the Internet. In addition, that application could be integrated into various librarian software systems. The contribution of this work is in the model architecture of the system for cataloguing in the UNIMARC format. User interface (described by use case diagrams) is connected with object model of UNIMARC format. According to that, any change or addition of new input data set of library documents requires only the change of use case diagrams which describe user interface while the rest of the model remains the same.