To achieve their potential therapeutic benefit, hyaluronic acid derivatives should be injected directly into the knee joint space and not into the anterior fat pad or the subsynovial tissues. In the absence of a knee effusion, reproducible needle placement into the intra-articular space presents a challenge to the clinician. The accuracy of needle placement was assessed in a prospective series of 240 consecutive injections in patients without clinical knee effusion. The injections were performed by one orthopaedic surgeon using a 2.0-in (5.1-cm) 21-gauge needle through three commonly employed knee joint portals: anteromedial, anterolateral, and lateral midpatellar. Accuracy rates for needle placement were confirmed with fluoroscopic imaging to document the dispersion pattern of injected contrast material. Of eighty injections performed through an anterolateral portal, fifty-seven were confirmed to have been placed in the intra-articular space on the first attempt (an accuracy rate of 71%). Sixty of eighty injections performed through an anteromedial approach were intra-articular on the first attempt (75% accuracy rate), as were seventy-four of eighty injections performed through a lateral midpatellar portal (93% accuracy rate). Using real-time fluoroscopic imaging with contrast material, we demonstrated the difficulty of accurately placing a needle into the intra-articular space of the knee when an effusion is not present. This study revealed that a lateral midpatellar injection (an injection into the patellofemoral joint) was intra-articular 93% of the time and was more accurate than injections performed by the same orthopaedic surgeon using either of the other two portals. This study highlights the need for clinicians to refine injection techniques for delivering intra-articular therapeutic substances that are intended to coat the articular surfaces of the knee joint.