Alzheimer's disease (AD) is characterized by fibrillar deposits of amyloid-β (Aβ) peptides and neurofibrillary tangles of Tau proteins. Aβ peptides are composed of 37 to 49 residues, in which the Aβ42 isoform is particularly toxic and aggregation-prone and is enriched in the plaques of AD brains and thus considered central to the development of AD. Therefore, disaggregation and disruption provide potential therapeutic approaches to reduce, inhibit, and even reverse Aβ aggregation. Here we capture the atomic-level details of the interactions between sigmoid Aβ42 fibril 2MXU/5KK3 and either natural tanshinone compounds TS1/TS0 or negative charged ER, proposing two unprecedented disassembly mechanisms. Natural TS1/TS0 prefers to insert into the cavity together with part at the surface of the 2MXU to open up the mouth and twist the conformation, destroying the ordered growth of subsequent monomers along the fibril axis. For the more compact two-fold 5KK3, attachment of TS1/TS0 at the surface including some inserted in cavity results in the separation of the two folds. In the two sigmoid fibril systems, it is no longer applicable for the routine criterions to assess Aβ42 fibril disassembly by introduction of these drugs, such as either reduced H-bond number, decreased β-sheet contents or both. ER, like-charged to Aβ42 fibril, is especially exceptional, and departs utterly from the neutral ones to disassemble Aβ42 fibril. Besides of the inapplicable routine criterions, positive binding energy between ER and Aβ42 fibril also deviates from the hypotheses of "ligands exhibiting greater affinity for the β-amyloid peptide are effective at altering its aggregation and inhibiting cell toxicity" (Cairo et al, Biochemistry 2002, 41, 8620 8629), but resulting in stronger disassembly effect on the two kinds of sigmoid Aβ42 fibrils than neutral TS0/TS1. The disassembly power of charged ER molecules derives from its stronger deformation ability to the conformation of Aβ42 fibril than the neutral ones, twisting the one-fold 2MXU in tapered-shape and separating two-fold 5KK3 in two parts more further, that is in great agreement with experimental observations (Irwin, et al. Biomacromolecules 2013, 14 (1), 264-274). The unusual disassembly mechanisms fill the gaps and offer an alternative direction in engineering new inhibitors to treat AD.