Abstract Objectives Correct determination of penicillin susceptibility is pivotal for using penicillin in the treatment of Staphylococcus aureus infections. This study examines the performance of MIC determination, disc diffusion and a range of confirmatory tests for detection of penicillin susceptibility in S. aureus. Methods A total of 286 consecutive penicillin-susceptible S. aureus blood culture isolates as well as a challenge set of 62 MSSA isolates were investigated for the presence of the blaZ gene by PCR and subjected to penicillin-susceptibility testing using broth microdilution MIC determination, disc diffusion including reading of the zone edge, two nitrocefin tests and the cloverleaf test. Results Using PCR-based detection of blaZ as the gold standard, both broth microdilution MIC testing and disc diffusion testing resulted in a relatively low accuracy (82%–93%) with a sensitivity ranging from 49%–93%. Among the confirmatory tests, the cloverleaf test performed with 100% accuracy, while zone edge interpretation and nitrocefin-based tests increased the sensitivity of β-lactamase detection to 96%–98% and 82%–96% when using MIC determination or disc diffusion as primary test, respectively. Conclusions This investigation showed that reliable and accurate detection of β-lactamase production in S. aureus can be obtained by MIC determination or penicillin disc diffusion followed by interpretation of the zone edge as a confirmatory test for apparently penicillin-susceptible isolates. The more cumbersome cloverleaf test can also be used. Nitrocefin-based tests should not be used as the only test for confirmation of a presumptive β-lactamase-negative isolate.