Data mining is the process of extracting the useful information from the large amount of available data. Association rule mining is a popular tool for discovering useful associations from large amount of data. Once private data is released for mining, it is very difficult to prevent its misuse. Useful associations with hidden information or knowledge that are sensitive to the database owner could be easily exposed using this kind of tool. Therefore it is necessary to hide all the sensitive information that can be mined from the data in the form of association rules before releasing the data. Most of the methods proposed in literature for association rule hiding deals with binary database and few methods for quantitative database suffer with side effects. This paper proposes an approach for hiding sensitive association rules using differential evolution and mutual information. The proposed algorithm hides the rule by decreasing Association Measure of the rule below threshold. Side effects are reduced by choosing the items with higher mutual information. Experimental results on real datasets demonstrate that the proposed method can effectively sanitize the sensitive data with fewer side effects to the non-sensitive data.