Lung cancer is becoming one of the leading causes of cancer mortality. According to WHO’s statistics, lung cancer patients can be increased to 32,99,640 within 2040–41. It can start in any part of lungs, but most of the cases found in epithelial cells. Also small cell types and non-small cell types are commonly found. Often cancer shows its major symptoms at an advanced stages when it has spread to a large area of lungs. That is why early detection is necessary to prevent the further spread and neural network-based systems has seen a feasible advancement to develop clinically acceptable solution with enhanced accuracy and sensitivity to enhance endurance rate of patients. This study is focused on developing an effective and efficient way to diagnose lung cancer using computer-assisted diagnostic method. Many image processing methods such as preprocessing, classification and segmentation are applied a large number of computer tomography (CT) image samples for analysis by introducing a novel strategical employment of an intelligent and hybrid deep learning algorithm, i.e., Deep Learning Approached Training and Prediction (DLATP). The study is contained with an appropriate combination of morphological techniques, Otsu’s threshold and renowned method of diving the Support Vector Machine (SVM) as classifier. The performance analysis of DLATP method is obtained based on its accuracy, F1-score, precision, and sensitivity. The average error rate is 0.0012738 which is very promising. The result shows that performance was very significant in reducing FPR and reducing false positives with an effective accuracy rate of 99.74%, at epoch 118 the gradient value is generated 0.075057.