We compared different methods for whole-brain and grey matter (GM) atrophy estimation (ANTs v1.9, CIVET v2.1, FSL-SIENA(X) v5.0.1, Icometrix-MSmetrix v1.7, and SPM v12) in multiple sclerosis (MS). The accuracy and precision were evaluated for cross-sectional and longitudinal whole-brain and GM atrophy measures. All software showed high accuracy and comparable repeatability for cross-sectional measures. However, since there was poor reproducibility and high variability in cross-sectional and longitudinal atrophy measures, changes of MR scanner should be avoided. This study may help in the selection of a suitable pipeline, depending on the requirements of the application (research center, clinical setting or clinical trial).