Emerging evidence have shown that the patients with Alzheimer's disease (AD) often have a higher risk of later developing type II diabetes (T2D), and vice versa, suggesting a potential pathological link between AD and T2D. Amyloid-β (Aβ) and human islet amyloid polypeptide (hIAPP) are the principle causative components responsible for the pathologies of AD and T2D, respectively. The cross-sequence interactions between Aβ and hIAPP may provide a molecular basis for better understanding the potential link between AD and T2D. Herein, we systematically modeled and simulated the cross-sequence aggregation process, molecular interactions, and polymorphic structures of full-length Aβ and hIAPP peptides using a combination of coarse-grained (CG) replica-exchange molecular dynamics (REMD) and all-atom molecular dynamics (MD) simulations, with particular focus on the effect of association models between Aβ and hIAPP on the structural stability and polymorphic populations of hybrid Aβ-hIAPP aggregates. Four distinct association models (double-layer, elongation, tail-tail, and block models) between Aβ and hIAPP oligomers were identified, and the associated polymorphic Aβ-hIAPP structures were determined as well. Among them, different association models led to different Aβ-hIAPP aggregates, with large differences in structural morphologies and populations, interacting interfaces, and underlying association forces. The computational models support the cross-sequence interactions between Aβ and hIAPP pentamers, which would lead to the complex hybrid Aβ-hIAPP assemblies. This computational work may also provide a different point of view to a better understanding of a potential link between AD and T2D.