Injection into the acromioclavicular (AC) joint is often inaccurate (approximately 50%) even in experienced hands. In light of new anatomic observations, we evaluate accuracy of an innovative ultrasound-guided method and follow the clinical course of successful therapeutic injections. Relevant anatomy was investigated in 200 three-dimensional computed tomography scans, 100 magnetic resonance images, and 14 cadavers. Baseline measurements of joint depth and width were performed ultrasonically in 100 normal volunteers; 50 symptomatic patients were injected. Uniquely in a clinical ultrasound study, injection success was documented by arthrography. Outcomes after concomitant steroid instillation were observed for 6 months by visual analog scale (VAS) scores and pain provocation test results. Anatomic studies showed that the widest area for joint penetration was anterior superior. Injection success rate was 96%, overwhelmingly on the first needle pass. Shallow joint depth allowed access with a standard 3-cm needle. Joint width diminished with age but did not reduce injection success. Cadaveric joints admitted 1.2 ± 0.5 mL, but fluid ingress was initially blocked by soft tissues in one third of both cadaveric and clinical cases. Diligent follow-up after steroid injection showed sustained pain relief in the majority with isolated AC disease but significantly less in those with concomitant shoulder disorders. This high level of clinical injection success, irrefutably substantiated with arthrography, has not been previously demonstrated. The anterior superior aspect of the joint is the preferred place for entry. Initial intra-articular blockage to fluid inflow is common but can be surmounted. Encouraging 6-month results of steroid instillation in isolated AC disease do not apply to patients with coexisting shoulder pathologic processes.