Recently, methods based on the analysis of arbitrarily amplified target sites of genome microorganisms have been extensively applied in microbiological studies, and especially in epidemiological studies. The range of their application is limited by problems with discrimination and reproducibility resulting from a lack of standardized and reliable methods of optimization. The aim of this study was to obtain optimal parameters of the Random Amplified Polymorphic DNA (RAPD) reaction by using an orthogonal array as per the Taguchi and Wu protocol, modified by Cobb and Clark for Candida parapsilosis isolates. High Simpson’s index values and low Dice coefficients obtained in this study indicated a high level of interspecies DNA polymorphism between C. parapsilosis strains, and the optimized RAPD method proved useful in the microbiological and epidemiological study.