Our goal was to analyze minimum inhibitory concentration (MIC) data for Staphylococcus aureus isolated from surgical infections (SIs) and to look for correlations among the clinically available antimicrobials that were tested. Clinical isolates from SIs were collected by a multicenter surveillance group involving 34 institutions in Japan. During the period April 1998 to March 2007, 312 strains of S. aureus [71 methicillin susceptible (MSSA) and 241 methicillin resistant (MRSA)] were consecutively obtained from these institutions. MIC data for 18 clinically available antimicrobial agents [ABPC, CEZ, CTM, CMX, CPR, FMOX, CFPM, CZOP, IPM, MEMP, GM, ABK, MINO, CLDM, FOM, LVFX, VCM, and TEIC (abbreviations defined in Tables 2 and 3)] against these isolates was analyzed using a principal component analysis (PCA). PCA revealed that four principal components explained 71.1% of the total variance. The first component consisted of major contributions from MEPM and IPM. The second component consisted of major contributions from MINO. These two-first axes, which were strong and explained 54.2% of the total variance, were able to classify the clinical isolates into four clusters. Furthermore, the proportion of the four clusters provided the characteristics of the S. aureus that were clinically isolated at each institute. PCA is a clinically applicable method for analyzing MIC patterns. Such analyses might contribute to the establishment of a practical classification of antimicrobial agents and to the identification of the characteristic antimicrobial resistance patterns at each institute.