The main aim of Thesis is to develop a fuzzy classification methodology for the use in processing and analysis of bioinformatics data. The use of fuzzy algorithms is suggested for the task of bioinformatics data analysis due to their use of linguistic terms that are easily comprehensible for humans. The task of bioinformatics-based diagnostics approached in Thesis is formalized as a classification task in data mining, where one has access to a predefined class (diagnosis) for each observed case. The algorithms and methods to be included in any step of the methodology are determined experimentally by testing on twenty five actual bioinformatics data sets. The study focused on investigating different methods and evaluating for suitability for bioinformatics data. The methodology steps that needed improvement were identified in experimental examination. Therefore in this Thesis are also proposed a membership function construction method that is based on clustering approach, a modification of the accepted classification algorithm FuzzyBexa by introducing the principle of rule-stretching that is used in FURIA as well as a novel rule fuzzification method that applies information from the constructed membership functions. The experimental validation of the fuzzy classification system that implements the proposed methodology proves its efficacy in the bioinformatics classification task. The developed methodology can be easily adapted for other fuzzy classification algorithms. The methodology can also be applied to other areas with the necessity to process and analyze similar data structures (tens of thousands of attributes and comparatively few, approximately hundred, records). The work contains 160 p. (include appendixes), 31 tables, 47 figures,6 appendixes and 133 referances