In the article diagnosis problem as the main issue has been investigated and as instance of diagnostics issues, medical diagnosis has been researched. Usable and unusable mathematical principles in the expert systems have been analyzed in the next step. Bayesian formula for calculate of occurrence probability of a hypothesis (assumption) has been surveyed. As well as utilization of this formula for calculating the probability of a disease is suggested with justifiable reasons. To show the function of these mathematical principles in reasoning cycle of each expert system, causal-covering method has been investigated. This method discloses the function of inference engine in expert systems. The determined diagnosis is actually comparative cognition.