Data clustering is one of the prominent fields of data mining which detects natural groups in a dataset. For the high dimensional dataset, traditional methods generally do not perform efficiently to cluster the data. Therefore, this paper proposes a novel metaheuristic method for data clustering based on k-means and improved cuckoo search to extend the capabilities of traditional clustering methods. The effectiveness of proposed method is tested on the three microarray datasets. Experimental results validate that the proposed method outperforms the existing methods.