Abstract: Data mining extracts valuable knowledge from large amounts of data. It is a powerful new technology to help companies focus on the most important information in their data warehouses. Several issues need to be addressed when mining on data is performed that are bulk at size and geographically distributed at various sites. Privacy preserving data mining has emerged as promising way to mining knowledge from large databases securely. The techniques are classified as: data distribution, data modification, and data mining algorithm, rule hiding and privacy preservation. Our approach is based on heuristics of Association Rule Mining. The techniques involves modifying database to prevent sensitive rules from getting disclosed which leads to information loss of non-sensitive data. We propose an algorithm that hides sensitive item on either side of rule by selective modification of database with minimal information loss. The prime objective is the accuracy of algorithm in terms of rule hiding, ghost rules and missing rules