Résumé : The threats of substandard and falsified (SF) antimicrobials, posed to public health, include serious adverse drug effects, treatment failures and even development of antimicrobial resistance. Next to these issues, it has no doubt that efficient methods for on-site screening are required to avoid that SF antimicrobials reach the patient or even infiltrate the legal supply chain. This study aims to develop a fast on-site screening method for SF antimicrobials using spectroscopic techniques (mid infrared, benchtop near infrared, portable near infrared and Raman spectroscopy) combined with chemometrics. 58 real-life illegal antimicrobials (claiming 18 different antimicrobials and one beta-lactamase inhibitor) confiscated by the Belgian Federal Agency for Medicines and Health Products (FAMHP) and 14 genuine antimicrobials were analyzed and used to build and validate models.Two types of models were developed and validated using supervised chemometric tools. One was used for the identification of the active pharmaceutical ingredients (APIs) by applying partial least squares-discriminant analysis (PLS-DA) and another one was used for the detection of non-compliant (overdosed or underdosed) samples by applying PLS-DA, k-nearest neighbors (k-NN) and soft independent modelling by class analogy (SIMCA). The best model capable of identifying amoxicillin and clavulanic acid (co-amoxiclav), azithromycin, co-trimoxazole and amoxicillin was based on the mid-infrared spectra with a correct classification rate (ccr) of 100%. The optimal model capable of detecting non-compliant samples within the combined group of amoxicillin and co-amoxiclav via SIMCA showed a ccr for the test set of 88% (7/8) using mid infrared or benchtop near infrared spectroscopy. The best model for detecting non-compliant samples within the group of amoxicillin via SIMCA was obtained using mid-infrared or Raman spectra, resulting in a ccr of 80% for the test set (4/5) and a ccr for calibration of 100%. For the group of co-amoxiclav, the optimal models showed a ccr of 100% for the detection of non-compliant samples by applying mid-infrared, benchtop near infrared or portable near infrared spectroscopy. Taken together, the obtained models, hyphenating spectroscopic techniques and chemometrics, enable to easily identify suspected SF antimicrobials and to differentiate non-compliant samples from compliant ones.