This paper provides a comprehensive survey of deep learning methods applied to graph data. It categorizes these methods into five main categories: graph recurrent neural networks (Graph RNNs), graph convolutional networks (GCNs), graph autoencoders (GAEs), graph reinforcement learning (Graph RL), and graph adversarial methods. The authors review the development history, model architectures, and training strategies of each category, highlighting their key characteristics and differences. They also discuss the challenges and solutions for applying traditional deep learning architectures to graphs, such as irregular structures, heterogeneity, large-scale graphs, and interdisciplinary knowledge integration. The paper further explores the applications of these methods and outlines potential future research directions.This paper provides a comprehensive survey of deep learning methods applied to graph data. It categorizes these methods into five main categories: graph recurrent neural networks (Graph RNNs), graph convolutional networks (GCNs), graph autoencoders (GAEs), graph reinforcement learning (Graph RL), and graph adversarial methods. The authors review the development history, model architectures, and training strategies of each category, highlighting their key characteristics and differences. They also discuss the challenges and solutions for applying traditional deep learning architectures to graphs, such as irregular structures, heterogeneity, large-scale graphs, and interdisciplinary knowledge integration. The paper further explores the applications of these methods and outlines potential future research directions.