Coronary CT Angiography (CCTA) is a critical diagnostic tool for detecting abnormalities in the Coronary blood arteries. Automated techniques are required to increase diagnostic speed and accuracy because manual CTA image interpretation is a labor-intensive, tedious task that increases the risk of misinterpretation. For this proposed work, we provide AngioResUNet++, a ground-breaking system to segmenting abnormal arteries and classifying that segmented coronary arteries from CCTA images. The system combines the DeepUNet++ model for efficiently segmenting coronary vessel with a pretrained ResNet-50 network for dependable classification. This system automates segmentation of heart blood vessels, extracting key features including texture, statistical, and geometric characteristics, and classifies the arteries into following categories: Normal, Blood-flow reduced Blocked and Narrowed. Unlike traditional methods, which often suffer from low accuracy and inefficient feature extraction, AngioResUNet++ provides an efficient solution that combines feature extraction and classification in a unified pipeline. Both segmentation and classification performance are improved by the deep learning method, which yielded a 99.71% classification accuracy and a 96.62% segmentation score. These results demonstrate significant improvements over conventional techniques, offering a faster, more reliable tool for diagnosing coronary artery diseases. The proposed system supports clinical decision-making, enabling timely interventions and improving patient outcomes.