This paper extends the sequential Bayesian persuasion model (Li and Norman, 2021; Wu, 2023) to infinite-period games, where a group of senders take unlimited opportunities one-by-one to reveal information to one decision maker. We first solve for the Perfect Bayesian Equilibrium outcomes through a geometric approach that features a simple better-than-full-revelation principle. Next, we seek for a refinement of PBE outcomes by proposing a so-called common-knowledge-of-common-ground&nbsp;assumption, which relies on senders’ recursive induction through higher-order belief hierarchy.&nbsp;<span>The justification of this notion is provided by showing the monotonicity of a&nbsp;</span><span>sequence of sets of beliefs which are used to characterize the reasonable outcomes&nbsp;</span><span>of the persuasion process. Under this refinement, we find a counter-intuitive result&nbsp;</span><span>that when the conflicts between two senders become more intense, the equilibrium&nbsp;</span><span>outcome can be less informative in such an infinite-period game. Finally, we make&nbsp;</span><span>comparison with finite-period persuasion games, and find that it is possible that, once&nbsp;</span><span>removing the ending period, senders achieve Pareto improvement.</span>