Background: In multiple sclerosis (MS), abnormalities of brain network dynamics and their relevance for cognitive impairment have never been investigated. Objectives: The aim of this study was to assess the dynamic resting state (RS) functional connectivity (FC) on 62 relapsing-remitting MS patients and 65 sex-matched healthy controls enrolled at 7 European sites. Methods: MS patients underwent clinical and cognitive evaluation. Between-group network FC differences were evaluated using a dynamic approach (based on sliding-window correlation analysis) and grouping correlation matrices into recurrent FC states. Results: Dynamic FC analysis revealed, in healthy controls and MS patients, three recurrent FC states: two characterized by strong intra- and inter-network connectivity and one characterized by weak inter-network connectivity (State 3). A total of 23 MS patients were cognitively impaired (CI). Compared to cognitively preserved (CP), CI-MS patients had reduced RS-FC between subcortical and default-mode networks in the low-connectivity State 3 and lower dwell time (i.e. time spent in a given state) in the high-connectivity State 2. CI-MS patients also exhibited a lower number and a less frequent switching between meta-states, as well as a smaller distance traveled through connectivity states. Conclusion: Time-varying RS-FC was markedly less dynamic in CI- versus CP-MS patients, suggesting that slow inter-network connectivity contributes to cognitive dysfunction in MS.