Convolutional neural network is a deep learning network, which is more suitable for image and speech processing. Image classification is one of the main research areas of visual artificial intelligence, with many applications, such as scene detection and robot vision enhancement. This paper proposes a new convolutional neural network to evaluate the potential of mangrove wetlands and discusses the development of the network in the film and television cultural and creative industries. Mangrove wetland is a special type of marine wetland centered on mangrove plant communities in tropical and subtropical coasts and intertidal zones of estuaries. They play an important role and play an important role in providing food, protecting wild germplasm resources, and population survival. Over the years, due to population pressure and economic development, the quality of mangrove wetlands has declined, the area has shrunk, and the restoration of mangrove wetlands has become a global hot spot. The development of the film and television industry is an important way to promote China’s cultural innovation. It is essential for strengthening the country’s soft power and improving China’s international image. However, due to the influence of many factors, the development of China’s film and television cultural creation industry is slow, and the development process is lagging behind. In recent years, the country has considered the development level of the cultural industry as an important development strategy, and has been taking various measures to promote the development of the film and television cultural industry. The creative film and television industry should be aware of its responsibility for cultural dissemination, actively learn from foreign industry development experience, and explore a scientific development system that meets China’s development needs and actual national conditions.