Attention Deficit/Hyperactivity Disorder (ADHD) often persists into adulthood. However, few screening tools have been adapted to assess adult ADHD using the DSM-5 criteria. This study assessed the diagnostic accuracy of a French version of the ADHD Self-Report Screening for DSM-5 (ASRS-5). This multicentric cross-sectional study included 557 participants: 309 adult ADHD outpatients without bipolar disorder (BD)/borderline personality disorder (BPD) (n = 236) or with BD/BPD (n = 36) and 285 adults without ADHD who were either healthy volunteers (n = 248) or outpatients with BD or BPD (n = 37). Measures included ADHD diagnosis and the ASRS-5. The ASRS-5 was a good predictor of ADHD diagnosis (cut-off score ≥ 13/24: sensitivity = 84.3%, specificity = 91.9%) in the sample of adult outpatients without comorbid disorders/healthy controls. Performances were lower with this cut-off score in some subgroups, notably low-severity ADHD symptomatology (sensitivity = 63.5%) and participants with BD or BPD (sensitivity = 91.7%, specificity = 54.1%). The French ASRS-5 had acceptable screening properties, even if its performance varied according to clinical variables. Further evidence is needed for patients with comorbid disorders having overlapping symptoms.