This paper describes a fuzzy approach to computer-aided medical diagnosis in a clinical context for dermatitis. This research will use Mamdani fuzzy inference system for determining Dermatitis risk level. The Diagnostic system of Dermatitis based on Fuzzy logic are constructed with seven indication variables. These variables have different intervals and used for determining status of domains in membership function of variables. Knowledge base in this system is constructed by production rules (IF-THEN). Fire Strength are obtained in each fuzzy rules base for each type of Dermatitis, then composite by using Max-Min method. The final result is an output namely the risk Dermatitis level.