Machine-learning techniques often encounter significant challenges when dealing with high-dimensional data due to the substantial memory requirements and extended processing times involved. In this paper, we introduce a novel unsupervised feature selection method called Entropy based Embedded Unsupervised Feature Selection (EEUFS). This method aims to select a suitable and concise subset of features while minimizing redundancy by assessing the entropy of each feature. Computing entropy for each feature, evaluating correlation between features and eliminating redundant ones, our approach reduces the feature set size while retaining essential information from the original dataset. Experimental evaluations conducted across six datasets using six different entropy measures demonstrate that our proposed Entropy based Embedded Unsupervised Feature Selection(EEUFS) method is effective and often outperforms state-of-the-art unsupervised feature selection methods.