Background and objectives: In hemophiliac patients, recurrent intra-articular bleeding leads to progressive arthropathy. Magnetic resonance imaging (MRI) is the gold-standard technique for joint assessment. The development of ultrasound protocols and standardized scales such as “Hemophilia Early Arthropathy Detection with Ultrasound” (HEAD-US) and “Point-of-care Ultrasonography” (POC-US) allow patient monitoring. Our main objective is to establish the diagnostic accuracy of ultrasound in the early detection of hemophilic arthropathy, with MRI as the gold standard. The interobserver variability in the ultrasound and MRI scales, the association between the POC-US and HEAD-US protocol, and the ability of ultrasound to detect haemosiderotic remnants will also be established. Patients and Methods: 45 hemophiliac patients were included. 180 joints (knees and ankles) were evaluated with MRI (Denver scale) and Ultrasound (HEAD-US and POC-US scale). Results: There is a high and statistically significant association between the ultrasound and MRI scales. Only for the ankle effusion variable the association was medium. The specificity of HEAD-US was ≥90% in both joints. Sensitivity in the knee did not exceed 82% in any assessed aspect, being high in the ankle for the cartilage and bone alterations (97%) but low for effusions (55%). In general, the predictive values were high, with the exception of the negative predictive value of joint leak in knee and ankle (close to 70%). Ultrasound did not detect haemosiderotic remnants in any case. Statistically significant interobserver agreement was obtained for HEAD-US and Denver-MRI assessment. The level of association between the HEAD-US scale and the POC-US protocol was medium. Conclusions: Ultrasound is a technique with high diagnostic accuracy and reproducibility to detect and quantify the early signs of hemophilic arthropathy. Ultrasound is insensitive to detect haemosiderin in synovial hypertrophy.