Objective The American College of Radiology (ACR) recently published the ovarian-adnexal reporting and data system (O-RADS) to provide guidelines to physicians who interpret ultrasound (US) examinations of adnexal masses (AM). This study aimed to compare the O-RADS with two other well-established US classification systems for diagnosis of AM. Methods This retrospective multicenter study between May 2016 and December 2019 assessed consecutive women with AM detected by the US. Five experienced consultant radiologists independently categorized each AM according to O-RADS, gynecologic imaging reporting and data system (GI-RADS), and international ovarian tumor analysis (IOTA) simple rules. Pathology and adequate follow-up were used as reference standards for calculating the validity of three US classification systems for diagnosis of AM. Kappa statistics were used to assess the inter-reviewer agreement (IRA). Results A total of 609 women (mean age, 48 ± 13.7 years; range, 18–72 years) with 647 AM were included. Of the 647 AM, 178 were malignant and 469 were benign. Malignancy rates were comparable to recommended rates by previous literature in O-RADS and IOTA, but higher in GI-RADS. O-RADS had significantly higher sensitivity for malignancy than GI-RAD and IOTA ( p = 0.003 and 0.0007, respectively), but non-significant slightly lower specificity ( p > 0.05). O-RADS, GI-RADS, and IOTA showed similar overall IRA ( κ = 0.77, 0.69, and 0.63, respectively) with a tendency toward higher IRA with O-RADS than with GI-RADS and IOTA. Conclusions O-RADS compares favorably with GI-RADS and IOTA. O-RADS had higher sensitivity than GI-RADS and IOTA simple rules with relatively similar specificity and reliability. Key Points • The malignancy rates were comparable to recommended rates by previous literature in O-RADS and IOTA, but higher in GI-RADS. • The O-RADS had significantly higher sensitivity for malignancy than GI-RADS and IOTA (96.8% vs 92.7% and 92.1%; p = 0.003 and 0.0007, respectively), but non-significant slightly lower specificity (92.8% vs 93.6% and 93.2%, respectively; p > 0.05). • The O-RADS, GI-RADS, and IOTA showed similar overall inter-reviewer agreement (IRA) (κ = 0.77, 0.69, and 0.63, respectively), with a tendency toward higher IRA with O-RADS than with GI-RADS and IOTA.