Metagenomics analysis has increased its importance in medicine with numerous recent research to investigate and explore the association of metagenomic data to human disease. Discretization approaches are proven as efficient tools to improve the disease predic- tion performance on metagenomic data. This study proposes a technique based on Entropy and combining some scaler algorithms to conduct bins for discretizing metagenomic data to perform disease classification tasks. Our disease prediction results on six bacterial species abundance metage- nomic datasets with the discretization method based on Entropy have revealed promising results compared to the Equal Width Binning with AUCs of 0.955, 0.826, 0.893, 0.692, 0.798, 0.765 classified by a One- dimensional Convolutional Neural Network on data including samples related to Liver Cirrhosis, Colorectal Cancer, Inflammatory Bowel Dis- ease (IBD), and two datasets of Type 2 Diabetes (namely, T2D, and WT2D), respectively.