The objective of this master’s thesis is to recognize and evaluate data mining algorithms which are commonly implemented in modern Medical Decision Support Systems (MDSS). They are helpful in a variety of healthcare units over the entire world. These institutions store enormous amount of medical data. This data may contain suitable medical information buried in different patterns concealed among the records. Inside the research some admired MDSS’s are analysed in order to make a decision for the most common data mining algorithms employed for a particular purpose by them. Three algorithms have been identified: Multilayer Perceptron, Naïve bayes and C4.5. A number of testing configurations are examined in order to make a decision of the best setting for the algorithms. After that, an eventual comparison of the algorithms orders them with respect to their manner of functioning. The assessment is rely on a set of performance metrics. In WEKA software analysis has been done and data has been taken from UCI repository medical datasets The data which is taken into consideration are breast cancer, heart disease, hepatitis . The analyses have shown that it is not very simple to name a single data mining algorithm to be the most suitable for the medical data. The consequences of data sets for the algorithms were very linked. However, the concluding evaluation of the outcomes allowed singling out the C4.5 to be the best classifier for the given domain. It was followed by the Naïve Bayes and the Multilayer Perceptron.