Abstract: In recent years, the state has increasingly emphasized the overall protection of regional and largescale cultural landscape heritages. However, due to the widespread distribution, vast area, and scattered nature of these heritages, traditional survey and analysis methods have difficulty in rapidly and accurately identifying them, leaving this field largely unexplored. This study takes the coastal areas of Jiangsu Province as the research area and attempts to utilize the improved DeepLabV3+ convolutional neural network from deep learning techniques to identify large-scale cultural landscapes. The aim is to construct a learning model that can achieve relatively accurate identification of large-scale cultural landscapes. Through this attempt, this paper hopes to delve deeper into the rich and valuable large-scale cultural landscape resources, as well as provides a new analytical perspective and method for the study of large-scale cultural landscapes, ultimately preserving, inheriting, and utilizing the diverse cultural heritage of China.