Aims. Ultrasound (US)-guided injections are widely used in patients with conditions of the shoulder in order to improve their accuracy. However, the clinical efficacy of US-guided injections compared with blind injections remains controversial. The aim of this study was to compare the accuracy and efficacy of US-guided compared with blind corticosteroid injections into the glenohumeral joint in patients with primary frozen shoulder (FS). Methods. Intra-articular corticosteroid injections were administered to 90 patients primary FS, who were randomly assigned to either an US-guided (n = 45) or a blind technique (n = 45), by a shoulder specialist. Immediately after injection, fluoroscopic images were obtained to assess the accuracy of the injection. The outcome was assessed using a visual analogue scale (VAS) for pain, the American Shoulder and Elbow Surgeons (ASES) score, the subjective shoulder value (SSV) and range of movement (ROM) for all patients at the time of presentation and at three, six, and 12 weeks after injection. Results. The accuracy of injection in the US and blind groups was 100% (45/45) and 71.1% (32/45), respectively; this difference was significant (p < 0.001). Both groups had significant improvements in VAS pain score, ASES score, SSV, forward flexion, abduction, external rotation, and internal rotation throughout follow-up until 12 weeks after injection (all p < 0.001). There were no significant differences between the VAS pain scores, the ASES score, the SSV and all ROMs between the two groups at the time points assessed (all p > 0.05). No injection-related adverse effects were noted in either group. Conclusion. We found no significant differences in pain and functional outcomes between the two groups, although an US-guided injection was associated with greater accuracy. Considering that it is both costly and time-consuming, an US-guided intra-articular injection of corticosteroid seems not always to be necessary in the treatment of FS as it gives similar outcomes as a blind injection. Cite this article: Bone Joint J 2021;103-B(2):353–359