The aim of our study was to assess the quality of antibiotic prescriptions in a sample of general practitioners (GPs) receiving junior doctors in training, whatever the motive of the prescription. We performed a prospective observational study of all antibiotics prescribed in October 2012 by 21 GPs working in southeastern France. Two specialists (general medicine and infectious diseases) independently assessed the compliance with recommendations of antibiotic prescriptions using a validated algorithm. Two hundred and thirty-two antibiotic courses were prescribed, mainly for low respiratory tract infections (30%), ENT (26%), urinary tract (22%) or skin (13%) infections. Forty prescriptions were considered as appropriate (17%), 77 as inappropriate (33%; mainly due to a non-recommended molecule choice [77%] or a too long treatment duration [44%]) and 115 prescriptions were unnecessary (50%), due to diagnostic issues. There were wide variations between GPs. An essential laboratory or imaging investigation was missing for 36% of prescriptions: chest X-ray for pneumonia (80% were missing), rapid antigen diagnostic test for acute pharyngitis (23% missing) and urine dipstick for urinary tract infections (80% missing). Fluoroquinolones and macrolides/synergistins accounted for 31% of the prescriptions, and were associated with a lower prevalence of appropriate prescriptions (7% and 2% respectively, P<0.001). There was a co-prescription of anti-inflammatory drugs in 15% of the cases. The misuse of antibiotics was frequent in this study. Improving the diagnostic workout is of paramount importance. Urgent actions are needed to improve antibiotic use in general practice.