Osteoarthritis (OA) of the centrodistal (CD) and tarsometatarsal (TMT) joints is a common cause of lameness in horses. Intra-articular diagnostic anaesthesia and/or therapeutic injection are relied upon to help diagnose and treat many horses with OA of these joints. The objective of this study was to determine the accuracy of arthrocentesis of the CD and TMT joints using a sample population of equine surgeons and surgery residents. Randomised experimental study. Six operators each injected four CD and four TMT joints in 12 sedated horses. The operators were randomly assigned to inject either the left CD and right TMT or the right CD and left TMT on four randomly assigned horses. The joints were injected with a 4 ml solution of contrast medium (2 ml), sterile saline (1.5 ml) and amikacin (0.5 ml). A minimum of two radiographs of each joint was obtained to evaluate the presence of contrast medium within the target joint. The TMT joint was successfully injected in 23/24 joints (96% accuracy). The CD joint was successfully injected in 10/24 joints (42% accuracy). Communication between the TMT and CD joints was visible in 26% of successful TMT injections. Communication between the CD and TMT joints was visible in 20% of successful CD injections. Despite specific requests to do so, we were unable to standardise the injection technique across all operators. The accuracy of injecting the TMT and CD joints of sedated horses was 96 and 42%, respectively. The CD joint was frequently missed with contrast medium being placed in the periarticular tissues. These data support the clinical impression of the difficulty of injecting the CD joint and suggests that practitioners should utilise ancillary methods, such as radiographs, to ensure proper needle placement.