Self-assembly of A beta peptides intd amyloid aggregates has been suggested as the major cause of Alzheimer's disease (AD). Nowadays, there is no medication for AD, but experimental data indicate that reversion of the process of amyloid aggregation reduces the symptoms of disease. In this paper, all 8000 tripeptides were studied for their ability to destroy A beta fibrils. The docking method and the more sophisticated MM-PBSA (molecular mechanics Poisson Boltzmann surface area) method were employed to calculate the binding affinity and mode of tripeptides to A beta fibrils. The ability of these peptides to depolymerize A beta fibrils was also investigated experimentally using atomic force microscopy and fluorescence spectroscopy (Thioflavin T assay). It was shown that tripeptides prefer to bind to hydrophobic regions of 6A beta(9-40) fibrils. Tripeptides WWW, WWP, WPW and PWW were found to be the most potent binders. In vitro experiments showed that tight-binding tripeptides have significant depolymerizing activities and their DC50 values determined from dose response curves were in micromolar range. The ability of nonbinding (GAM, AAM) and weak-binding (IVL and VLA) tripeptides to destroy A beta fibrils was negligible. In vitro data of tripeptide depolymerizing activities support the predictions obtained by molecular docking and all-atom simulation methods. Our results suggest that presence of multiple complexes of heterocycles forming by tryptophan and proline residues in tripeptides is crucial for their tight binding to A beta fibrils as well as for extensive fibril depolymerization. We recommend PWW for further studies as it has the lowest experimental binding constant.