The dissertation explores the role of key words in legal information retrieval and analyses the use of terminological dictionary – Eurovoc Thesaurus. Based on the research conducted, and by studying the legislative procedure, as well as the availability of legal regulations, the dissertation proposes a model of interrelated regulations which would in the future form a centralised system for storage, retrieval, archiving and long-term preservation of legal regulations. The dissertation consists of eight chapters: 1) Introduction, 2) Access to documents and official sources of information, 3) Description of the Croatian legislative procedure, 4) Legal databases – examining the data structure, connectivity and availability, 5) The collection of legal regulations and the creation of a digital archive in Croatian state administrative organization – Central State Office for the Development of Digital Society, 6) Documentation processing of legal regulations and use of Eurovoc Thesaurus, 7) Research results: key words and the use of Eurovoc, 8) Recommendations and proposal of a centralised system model for long-term storage of interrelated regulations. The theoretical part of the dissertation provides an overview of the legislative basis on the right of access to and reuse of public sector information, which prescribes the obligation for public authorities to publish laws and other regulations on their websites in machine-readable format which is easily searchable. The problem appears if the obligation is not implemented, or when regulations are published on multiple locations by several authorities. Based on literature review, the legislative procedure is described in detail, including drafting, adoption, and publication of legal regulations in the Official Gazette of the Republic of Croatia. The dissertation includes a comparative analysis of five legal databases. These were tested based on search examples and evaluated with regards to the individual functionalities of each database. Firstly, search engines were tested by performing a common simple search. The analysis of search results was conducted, with scope and accuracy being most relevant, followed by user-friendliness and additional functionalities. After making general remarks concerning each search engine, a checklist was developed to examine functionality of each search engine. In the theoretical part of the dissertation the author describes in detail the procedure for collecting and processing legal regulations at the Central State Office for the Development of Digital Society, as well as the digital archive that enables their long-term preservation. The use of Eurovoc, as a documentation tool for indexing, was examined in processing of legal regulations. This was done through automatic indexing projects and by using Eurovoc descriptors. In addition, the results of the CADIAL project are presented. The project was initiated in cooperation with the Katholieke Universiteit Leuven Interdisciplinary Centre voor recht en Informatica (ICRI) research team, experts from the University of Zagreb and the Joint Research Centre. It resulted in a publicly available CADIAL search engine that allows users permanent and easy access to the entire national legislation. Also, a good practice example of using Eurovoc in indexing EU's official documents on the EUR-Lex portal was analysed. As part of the main research goal, the aim was to determine and analyse the role of key words in the use of a specific search engine. In order to analyse user behaviour of CADIAL search engine users, and to get a better insight into existing search options, as well as the role of Eurovoc Thesaurus, user logs were analysed over a three-month period. The analysis included 16,270 user queries, which were then presented through five case studies, depending on the entry of the query in the search engine. The results of the analysis showed that most users perform searches according to the text and the title of regulations. Since the goal was to explore the role of key words while using search engines, it is discouraging that only 9.27% of users used the Eurovoc Thesaurus. Additionally, several semantic queries were examined more closely with regards to how well they match with the basic search functions. The aim was to find how many such queries correspond to the words in the title of the regulation chosen by the user, as well as to the words in the text of the regulation, and if they match with Eurovoc descriptors. The queries matched best with the words in the text, whilst the match was lowest with regards to the Eurovoc descriptors, which was to be expected considering the use of Eurovoc in search. As part of the research, a survey was conducted on user satisfaction with the CADIAL search engine. The aim of the survey was to examine the user satisfaction with the search engine, and it was conducted as a pop-up survey. Over a period of three months, 875 responses were analysed, with over 90% of users replying that they found the required regulation and as many as 94.77% of users plan to use the search engine again. The survey also showed that more than half of the users did not use Eurovoc Thesaurus during their search. The third segment represents an analysis of data and information on user behaviour collected by the Google Analytics platform. The aim of the data review was to analyse the data on the use of CADIAL search engine and regulations that users retrieved during a five-year period. Google Analytics has shown, among others, that the number of CADIAL users has doubled during the period of five years, and that users are increasingly accessing the search engine from their mobile devices. Based on the findings of the research, the author proposes recommendations and a centralised system model for long-term preservation of interrelated regulations. These include the necessary improvements of the existing legal search engine, integration with other systems to enable users a broader context of each individual legislative act, and the development of additional modules such as the European Legislation Identifier and a module for the search of consolidated texts of regulations. Recommendations and development of a centralised system based on the proposed model for the long-term preservation of interrelated regulations would result in a transparent and trusted system which allows users to search for legal regulations quickly, easily and free of charge.