In light of the growing demand to screen numerous potential cases of lung-related diseases and COVID-19, along with the challenge of high false negatives in PCR tests, there arises a requirement for an alternative and simple screening method that utilizes radiological images like chest X-rays. Simple machine learning (ML) and neural networks models offer swift and automated approaches to identify abnormalities and capture significant vectors from altered lung tissue, potentially indicating specific characteristics of various lung diseases and the COVID-19 virus. In this study, we employed advanced convolutional neural networks (CNNs) that are VGG-16, ResNet-152, and Custom CNN models to construct robust models for classifying lung deceases. This research used Chest X-ray 14 dataset training the model. Apart from the VGG-16, ResNet-152, and Custom CNN models, the Custom CNN model provided better accuracy than others. It has secured 91%, 89.7%, and 89.5% of training, testing, and validation accuracies on the selected data.