Fuzzy Set Theory has used many auxiliary methods in the trials of solutions of some medical problems. One of the attempts was the finding of the optimal level of drug action in the case when the clinical symptoms disappeared completely after the treatment (Gerstenkorn and Rakus, 1994; Rakus, 1991). However, there can occur such a morbid process in which the symptoms do not disappear after the treatment. The medication improves too high or too low level of the quantitative symptom but it still indicates the presence of the disease. It sometimes makes some problems to choose the medicine, which acts best because it can happen that most of them influence the same symptoms while they do not improve the others. A fuzzy decision making model tries to make easier to find such a drug which affects most of the symptoms in the highest degree. To solve this problem we propose an application of discrete values of the membership degrees in the model instead of the continuous ones that were tested in the paper of Rakus-Andersson and Gerstenkorn (1997). It should improve the thoroughness of the method and heighten the reliability of the accepted decision. It is also considered how to choose the best medicine in the circumstances when some decision-makers have different opinions about the priority of the tested drugs.