Finding cancer early often means better treatment results and a higher chance of life. Typically, malignancies that are detected in their early stages are more susceptible to treatment and have a greater likelihood of being completely cured. Tissue histology is often utilized for early detection; however, it is usually performed manually by pathologists, which may be time-consuming and prone to mistakes. This research addresses these issues by developing and deploying a computer-aided method for identifying lung cancer in whole-tissue slides. The study presents a novel automated approach for interpreting histopathological lung and colorectal images by utilizing ensemble Convolutional Neural Networks (CNNs) to detect zones of lung and colorectal cancer. This work highlights the efficacy of ensemble models in analyzing histopathological pictures and their capacity to improve the identification of lung and colon cancer. Histopathology data from the lungs and stomach (LC25000) are used to test the model. This finding establishes a foundation for forthcoming investigations and progress in automated medical diagnosis, which will confer benefits to patients and healthcare professionals as technology and machine learning further evolve. The assessment of the model's performance relies on criteria such as recall, precision, F1 score, and accuracy.