Abstract: In association rule mining, leakage of sensitive data can cause potential threats to privacy and data protection. In distributed database architecture, performing association rule mining following traditional privacy preserving techniques are not feasible. We present a novel privacy preserving association rule mining algorithm that uses cryptosystem technique to maintain privacy. We use FDM technique to find frequent itemsets. The support count is encrypted using RSA algorithm and forwarded to other sites. We use one data initiator, one data combiner and other parties as client in ARM process. Experimental results show that this method is flexible and ensures privacy during global support count calculation process.