The potential for the use of the disk diffusion method to accurately predict penicillin MICs for Streptococcus pneumoniae was investigated with penicillin (6 μg), methicillin (5 μg), and oxacillin (1 μg) disks. A total of 183 S. pneumoniae isolates were tested by three MIC procedures (agar dilution, microdilution, and E-test). Regression analyses of the geometric mean of the three MIC results against (i) the sum of the zone diameters for methicillin, penicillin, and oxacillin disks; (ii) the sum of the zone diameters for methicillin and penicillin disks; and (iii) each of the three individual zone diameters were performed. Calculated MICs were determined from each of these regression analyses and compared to the mean reference MICs. A high level of correlation was obtained with both the two- and the three-disk procedures (r = 0.97), with essential agreement rates (±1 doubling dilution) between MICs calculated by the three-disk procedure and the two-disk procedure and the mean reference MICs of 98.4 and 98.9%, respectively. No major or very major errors were obtained with the two- or three-disk procedures. The accuracy of the disks used individually was lower (r = 0.84 to 0.93). However, oxacillin and methicillin disk testing remain excellent for screening strains, with all penicillin-susceptible strains having zones of >21 and >22 mm, respectively. The combination disk procedure, which involves the use of three disks (methicillin, oxacillin, and penicillin) or two disks (methicillin and penicillin) for testingS. pneumoniae, can provide accurate penicillin MICs and qualitative category results that are comparable to results obtained by the E-test, agar, and microdilution MIC methods.