The lungs are the primary organs of the respiratory system, and an infection in the lungs can lead to significant illness. Pneumonia is a prevalent lung infection that may result from viral or bacterial pathogens. Pneumonia is often assessed at the clinical level using chest X-ray, with the results then analyzed to verify the infection and its severity. This project seeks to provide a MobileNet (MN)-based methodology for classifying the selected X-ray database into healthy and pneumonia categories. The developed approach encompasses many phases: Image collection and resizing, feature extraction utilizing the MN, feature reduction through the Butterfly Algorithm (BA), and classification accompanied by threefold cross-validation to assess the efficacy of the proposed system. This study use both individual and merged MN features for investigation. The proposed study demonstrates that the individual-feature strategy achieves accuracy of >92%, while the fused feature methodology attains a detection accuracy of >97% when employing the Decision-Tree (DT) classifier. These findings validate that the executed scheme yields a clinically relevant outcome.