Brain networks of Alzheimer’s disease (AD) represent the fluctuating functional connectivity with time. Using the sliding-window method, this study identified two different functional connectivity states at the large-scale network level in AD spectrum. The AD patients took more time in the tense-connected state II than the non-AD populations. Moreover, lipid metabolism-related factors affected the dynamic network connectivity across the AD spectrum populations. These results provide insights into the neural biological underpinnings of the dynamic networks reorganization for AD pathophysiology.