Abstract Multi-agent systems (MASs) are vulnerable to privacy disclosure when performing time-vary formation in public cyberspace through information exchange between agents. This paper investigates the privacy-preserving time-varying formation control under multi-sensor multi-rate sampling mechanism for MASs. Firstly, since leader state measurement is not entirely available in cyberspace for all followers, a distributed adaptive leader observer is constructed for each follower to estimate the state and parameter matrix of leader; Secondly, the information exchange strategies is designed based on Paillier encryption algorithm between followers to prevent their privacy from being leaked to internal and external eavesdroppers; Finally, due to the introduction of a multi-sensor and multi-rate sampling mechanism, some sub-states of the agent are discontinuous, a time-varying formation control scheme with Privacy-Preserving is designed by constructing dynamic estimator for each follower. Numerical simulation is carried out to verify the effectiveness of the proposed method.