In recent years, the use of methods to investigate muscle–tendon unit function that combine motion capture with ultrasound (MoCapUS) has increased. Although several limitations and individual errors of these methods have been reported, the total error from all the potential sources together has not been estimated. The aim of this study was to establish the total error in the Achilles tendon (AT) measurements, specifically its length (ATL), strain (ATS), and moment arm (ATMA) acquired with MoCapUS during running. The total error from digitizing, marker movement, ultrasound calibration, and probe rotation errors caused mean ATL error of 4.2 ± 0.6 mm, mean ATMA error of 0.1 ± 0.1 mm, and could potentially alter measured ATS by a mean 2.9 ± 0.2%. Correcting both the calcaneus insertion position (CIP) and properly synchronizing ultrasound and motion capture data caused changes of up to 5.4 ± 1.7 mm in ATL and 11.6 ± 1.3 mm in ATMA. CIP correction and synchronization caused a similar amount of change in ATL, as well as ATS. However, the ATMA change was almost exclusively due to the CIP correction. Finally, if all sources of error were combined, the total ATL error could reach 13.1 mm, the total ATMA error could reach 14.4 mm, and ATS differences could reach up to ± 6.7%. The magnitude of such errors emphasizes the fact that MoCapUS-based AT measurements must be interpreted within the scope of their corresponding errors.