This survey provides a comprehensive overview of recent advancements in 3D generation, covering fundamental methodologies, scene representations, generation methods, datasets, and applications. The field has seen significant progress, driven by the success of generative AI in image and video synthesis. Key methods include 3D-GAN, DeepSDF, DMTet, EG3D, DreamFusion, PointE, Zero-1-to-3, and Instant3D. The survey categorizes 3D generation into four algorithmic paradigms: feedforward generation, optimization-based generation, procedural generation, and generative novel view synthesis. It discusses various scene representations, including explicit (point clouds, meshes, multi-layer representations), implicit (Neural Radiance Fields, Neural Implicit Surfaces), and hybrid representations. The survey also covers datasets, applications such as 3D human and face generation, and open challenges in 3D content generation. It highlights the importance of differentiable rendering, neural scene representations, and the integration of generative models with scene representations. The survey emphasizes the need for innovative approaches to address the complexities of 3D generation, including the challenges of data scarcity, evaluation, and multi-view consistency. The survey concludes with a structured roadmap of recent works in 3D generation, aiming to provide a systematic summary for researchers and practitioners in the field.This survey provides a comprehensive overview of recent advancements in 3D generation, covering fundamental methodologies, scene representations, generation methods, datasets, and applications. The field has seen significant progress, driven by the success of generative AI in image and video synthesis. Key methods include 3D-GAN, DeepSDF, DMTet, EG3D, DreamFusion, PointE, Zero-1-to-3, and Instant3D. The survey categorizes 3D generation into four algorithmic paradigms: feedforward generation, optimization-based generation, procedural generation, and generative novel view synthesis. It discusses various scene representations, including explicit (point clouds, meshes, multi-layer representations), implicit (Neural Radiance Fields, Neural Implicit Surfaces), and hybrid representations. The survey also covers datasets, applications such as 3D human and face generation, and open challenges in 3D content generation. It highlights the importance of differentiable rendering, neural scene representations, and the integration of generative models with scene representations. The survey emphasizes the need for innovative approaches to address the complexities of 3D generation, including the challenges of data scarcity, evaluation, and multi-view consistency. The survey concludes with a structured roadmap of recent works in 3D generation, aiming to provide a systematic summary for researchers and practitioners in the field.