Abstract: In this study, we proposed a novel hybrid protection scheme to protect privacy information and clustering knowledge in mined data. The scheme involves integrating the privacy-preserving data mining technique with that of the knowledge-preserving anti-data mining technique. Hierarchical clustering is used and the clustering structure is manipulated by perturbation to create original data where the mined data has the appearance of similar information and knowledge from the original dataset but with misleading and non-useful contents. The scheme is novel in that it allows users to tailor the amount of protection on personal basis. Experimental tests were conducted on ten public datasets results showed that the privacy information in datasets is preserved and the clustering knowledge cannot be revealed in the mining process. Furthermore, the original dataset can be restored using the key values in the reverse order of the two phases perturbation procedures.