Precision of cleft lip and/or palate antenatal diagnosis plays a significant role in counselling, neonatal care, surgical strategies and psychological support of the family. This study aims to measure the accuracy of antenatal diagnosis in our institution and the detection rate of cleft lip and/or palate on routine morphologic ultrasonography. In this retrospective observational study, we compared antenatal and postnatal diagnosis of 233 patients followed in our unit. We classified our patients according to the Kernahan and Stark's classification system: Group 1: facial cleft including labial and labio-maxillary clefts; Group 2: facial cleft including total, subtotal and submucous palatal clefts; Group 3: labio-maxillary-palatal clefts. Out of 233 patients, 104 were antenatally diagnosed with a facial cleft, i.e., an overall detection rate of 44.6%. The diagnosis was confirmed at birth in 65 of these patients, i.e., an overall accuracy of 62.5%. Of the 67 children (29.2%) in Group 1, the screening detection rate was 58.2% with an antenatal diagnostic accuracy of 48.7%. Of the 97 children (41.6%) in Group 2, the screening detection rate was 2% with an antenatal diagnostic accuracy of 50%. Of the 69 children (29.6%) in Group 3, the screening detection rate was 91.3% with an antenatal diagnostic accuracy of 71.4%. Our study demonstrates a relatively poor diagnostic accuracy in prenatal ultrasound, where the diagnosis was inaccurate in one third to one half of patients. It showed great variability in the screening detection rate depending on the diagnostic group observed, as well as a low rate of detection of palatal clefts.