Manufacturers of geomatics instruments have designed a number of centering devices to perform appropriate positioning over surveying and geodetic points, and to reproduce these positions as many times as needed. Depending on the accuracy requirements of metrological, geodetic or topographic work, a choice can be made between conventional-centering or forced-centering, taking in to account commonly accepted deviations for each type of centering device. Miscentering values are currently stablished in two ways: through "experience" or from tests involving angular and linear measurements performed by surveying instruments placed on the tested centering devices. None of these methods is adequate enough to determine the true capabilities of centering systems. This gap in the analysis of a relevant field in high accuracy geomatics projects and metrological testing of geomatics instruments has encouraged this study. A new method for assessing the accuracy of centering devices is presented. This method avoids surveying measurements and takes advantage of the capabilities of two high-precision measuring devices—coordinate measuring machine (CMM) and measuring arm—as reference instruments, which ensure that the results have been determined in a fairly reliable way. The procedures performed with this new method applied to four centering devices and the corresponding results are exposed.