Through a multicenter prospective AIS cohort study, we analyze the potential association of human proteome, microbiome and metabolome alterations with AIS prognosis, searching for key proteins, differential organisms and metabolites, combining experimental data at multiple molecular levels with computational models, and establishing early prediction models through machine learning-based prediction algorithms. While closely tracking the recurrence of stroke in AIS patients, we evaluate the predictive value of human proteome, microbiome and metabolites for stroke recurrence through a nested case-control study, which provides key reference information for exploring the unknown residual risk of AIS recurrence.