Disease prediction is a rapidly growing research field in the age of information and technology. In the future, machinery robotics is expected to significantly support doctors in specific diagnostic areas and provide helpful medical advice. Metagenomics has been emerging in recent years and plays an essential role in personalized medicine approaches. However, data sources for metagenomic analysis are still complex and very high dimensional; hence it is necessary to provide helpful feature selection methods to apply to them. In this study, we propose to use the Chi-Squared technique to extract insights from the metagenomic features for disease diagnosis. Chi-Squared is deployed in feature selection on Metagenomic Data to extract meaningful features related to Inflammatory Bowel Disease and shows a somewhat spirited performance than other methods. As obtained results, Chi-Squared is expected to pick out valuable features that can improve the prediction performance.