In this work we present the usage of semantic web knowledge representation formalism (OWL) in combination with general purpose reasoning for building a medical expert system. The approach is illustrated by the example of the knowledge base construction for decision support tasks in the heart failure domain. The work includes presentation of descriptive knowledge in the ontological form and its integration with the heart failure procedural knowledge. A methodology for formalizing procedural knowledge using OWL semantics is presented and compared with the traditional rule-based expert systems approach. Main advantages of the approach are straightforward integration of knowledge components and knowledge maintainability. Main drawback is reduced formalism expressivity, which in some cases does not necessarily impose insuperable obstacle. Keywords. knowledge management, knowledge bases, semantic web, expert systems, heart failure, reasoning