Curcumin and related compounds have been validated to remove even well-developed human β-amyloid plaques from the brain of transgenic mice, in vivo. However, their molecular mechanism of the plaque buster activity is rather unknown. Computational chemistry was employed here to better understand the β-amyloid protein elimination. According to our docking studies, a tautomeric "keto-enol" flip-flop mechanism is proposed that may chop up β-amyloid plaques in Alzheimer's due to removing each hairpin-foldamers one by one from both ends of aggregated fibrils. According to the experimented models, other bi-stable "keto-enol" pharmacophores might be identified to break up amyloid plaques and enhance rapid clearance of toxic aggregates in Alzheimer's disease.