Every building diagnostic analysis begins with a detailed visual check of deteriorations of building elements. During this process the identification of touched components and the circumstances of the given failure are described. A properly completed survey determines the way of further analyses and interventions to be done. Thus the visual check may support indirectly the decisions which deal with the life cycle of the examined building or its part. In the case of ad hoc and independent surveys the decision-supporting ability of the traditional visual examination procedures proves to be inconsiderable, if the subjects of the surveys belong to the same property. Due to several subjective factors the evaluation of these surveys and their comparative analysis are almost impossible. It is supposable that the fuzzy expert system proposed in this article may remedy the problem mentioned above, owing to the peculiarities of this professional field (application of a variety of linguistic variables, uncertainty originated from different sources, assumptions based on experiences and observations). This paper sketches out the facts of uncertainties of the visual examination procedure in building diagnostics, then describes the architecture and components of the fuzzy expert system. Thereafter, the operation of this method is illustrated through three building diagnostics case studies. © 2009 IEEE.