<p>Lung tumor is a condition resulting from uncontrolled cell division in the lungs, where cells multiply as part of their normal function. It is difficult to detect because it arises and shows symptoms in the final stage. There are no prevention techniques for lung cancer, but early detection and diagnosis are critical in determining the chances of survival. However, various existing systems are utilized to detect lung tumors, but those approaches produce inaccurate results because the image quality is very low and the tumor portion is not segmented properly. To avoid this issues, the hybrid deep learning algorithm with an optimized pre-processing approach is developed to detect lung tumors. In the beginning, lung tumor images are collected from the CT scans. After being gathered, the images are pre-processed to enhance the original image's quality. For that purpose, Fast-Flexible Denoising Networks (FFD Net) are utilized to remove noise from the original image, while improved Reformed Histogram Equalization is used to raise the contrast level of the noise-removal image. In order to enhance the performance of the RHE, the Tunable parameter is optimally selected by using Tasmanian Devil Optimization (TDO). Trustworthy Multi-view Segmentation Network (TMS Net) is utilized to segment the tumor portion from pre-processed images. Subsequently, the features are gathered from the segmented images with the help of locally encoded transform feature histogram, multiscale Gabor rotation invariant, and colour correlogram. To detect the lung tumors, the extracted features are provided to a hybrid Multiunit Discrete Hopfield Neural Network (MDHNN) Multi-Layer Perceptron (MLP). The suggested method provides 93.40% of kappa, 92.73% of F1 score, 89.30% of MK, and 94.60% of accuracy. The designed model is the best choice for automated lung tumor detection because the disease detected at early stage and improve the living standard of the patients.</p>