Abstract: An improved harmony search algorithm (IHS) was proposed and used in clustering analysis. The feedback mechanism was introduced first into the harmony search algorithm, and harmony memory considering rate and bandwidth can be dynamically adjusted by calculating the difference between the best harmony and the worst harmony, which makes the IHS converge to the global optimal solution quickly. The proposed algorithm updates the decision variables which represent the cluster centers to minimize the value of objective function and get the best partition of data samples. A method which can automatically determine the true number of clusters of data samples was then proposed, the true number of clusters was calculated when the input number of cluster center is greater than the true number of clusters. Finally, four performance measures were used to compare the proposed algorithm with ANT-based and traditional harmony search (HS) based clustering analysis algorithms, it is shown that the performance of the proposed algorithm is better than the other two algorithms.