Histopathological images are essential for diagnosing medical conditions, especially when assessing a variety of illnesses, such as cancer. This research applied deep learning models to developed an automated lung cancer detection system. In this study we acquired dataset from kaggle, dataset consisting of 3000 images categorized into three different classes: adenocarcinoma, squamous carcinoma, and benign. Seventy percent (2100) images were used for training, while the remaining 15% (450) images each were allotted for testing and validation. After preprocessing, the initial image size of 768 × 768 was reduced to 224 × 224. The system achieved an impressive 98% testing accuracy using the VGG19 model, indicating its suitability for automated lung cancer finding and diagnosis. The VGG16 model reached with 95.33% accuracy, NasNetLarge demonstrated stable performance across 50 epochs with 91.77% accuracy, and InceptionV3 stabilized after 35 epochs, achieving 96.22% accuracy. The evaluation included precision, F1-score, and recall metrics for each class, providing a complete understanding of model performance. In overall summary, this analysis demonstrates the effectiveness of deep learning models in classifying lung cancer histopathology images for disease identification, VGG19 can be effective tools for automated identification of lung cancer, with high accuracy and excellent result. The difference between the highest accuracy achieved by VGG19 98% and the lowest accuracy achieved by NasNetLarge 91.77% is 6.23%.