Radiostereometric analysis (RSA) is a method for measuring micromotion in joint arthroplasties. RSA has never been used in total wrist arthroplasties. We evaluated: (i) the precision of model-based RSA in total wrist arthroplasties measured in a phantom model and in patients; (ii) the number of bone markers necessary to ensure the precision; and (iii) the accuracy of model-based RSA in a phantom model. Reverse engineered models of radial and carpal/metacarpal components of two wrist arthroplasties (ReMotion® and Motec®) were obtained by laser scanning. Precision and accuracy of each arthroplasty were analyzed with regards to translation and rotation along the three coordinate axes. Precision was analyzed in 10 phantom and 30 clinical double examinations for each arthroplasty, and was expressed by a repeatability coefficient. The precision of different numbers and configurations of bone markers in the phantom model were compared. Accuracy was tested in a phantom model where the implants were attached to a micrometer, and was defined as the mean difference between measured and true migration. In the phantom model the precision for translations ranged from 0.03 to 0.14 mm and for rotations from 0.18 to 1.52°. In patients the precision for translations ranged from 0.06 to 0.18 mm, and for rotations from 0.32 to 2.18°. Less than four bone markers resulted in inferior precision. Accuracy ranged from -0.06 to 0.04 mm, and from -0.38 to -0.01°. Y-rotations could not be obtained from the Motec® due to rotational symmetry about the longitudinal axis. We conclude that model-based RSA in total wrist arthroplasties is precise, accurate, and feasible to use for clinical evaluation of micromotion in wrist arthroplasties. © 2018 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 9999: 1-12, 2018.