The process of selecting most suitable optimal features plays important role in improving both performance and interpretability of classification models. This paper introduces a fusion approach, referred to as BFO-ML (Bacterial Foraging Optimization - Machine Learning), which combines the capabilities of Bacterial Foraging Optimization (BFO) with modern machine learning techniques. The goal of BFO-ML is to streamline feature selection process by utilizing foraging behavior of bacteria to systematically navigate complex feature space. By exploring feature subsets, BFO-ML identifies and optimizes those subsets that contribute much to classification accuracy. This integrated approach is designed to enhance feature selection process improving classification accuracy. The experimentation and evaluation, reveal that the proposed BFO-ML based model for feature selection and classification is more efficient than the existing methods.