Due to the enhancement in the quantity of data across the world, it becomes very complex to analyze those data, Categorize those data into singular collection. This may leads to the requirement for better data mining techniques. Clustering is the one of the best data mining technique. One of the mostly used clustering techniques is k-means clustering. It is very simple and effective for clustering. This paper provides a new technique to enhance the k-means clustering, which can result in better performance. For initialization, this paper uses an improved version of Hopfield Artificial Neural Network (HANN) algorithm.