The paper presents HADES - Hypermedia Adaptive Diagnostic Expert System. It can be used for diagnosing various farm machinery. The requirements, overview and analysis of the existing state-of-the-art with respect to the diagnostic expert systems are presented. The general architecture of HADES is described, also the methods used for knowledge representation and knowledge acquisition. The choice of a combination of case-based, model-based, and rule-based reasoning methods is substantiated. A diagnostic problem solving strategy is suggested and realized as a prototype for diagnosing brake and fuel systems. The main features and advantages of the prototype are described in the paper.