This study describes a deep learning technique based on the U-Net architecture for accurately segmenting lung, trachea, and heart tissue from chest CT data. Chest CT imaging is an important diagnostic technique in radiology, and correct segmentation of lung structures is required for the diagnosis and treatment of a variety of disorders. The U-Net model was trained and tested on a dataset of 3342 CT images and associated segmentation masks. The model was trained to recognize the complex patterns and characteristics of lung, trachea, and heart tissue, allowing it to execute precise segmentation. During the validation phase, the model performed admirably, as demonstrated by dice and Jaccard scores of 0.887 and 0.861, respectively. These scores demonstrate the model's ability to reliably recognize nodules and capture the borders and contours of lung tissue. The proposed method has a lot of promise for use in clinical settings. It can help radiologists diagnose and treat lung, trachea, and heart diseases by automating the segmentation process. The model's exact delineation of lung tissue and nodules can help in the diagnosis and characterization of anomalies, allowing radiologists to make educated decisions about patient management. Implementing this strategy in healthcare workflows can have several advantages. It minimizes the amount of manual work necessary for segmentation, resulting in increased efficiency and perhaps shorter turnaround times in radiological evaluations. Furthermore, precise segmentation of lung, trachea, and heart tissues allows for quantitative investigation of disease progression and therapy response evaluation. It also aids in the planning of surgical operations, assisting clinicians in developing effective treatment methods.