This paper provides a comprehensive overview of deep meta-learning based recommendation systems. It discusses the challenges faced by deep learning-based recommendation methods due to data sparsity and computational inefficiency. The paper introduces meta-learning as a paradigm to improve learning efficiency and generalization ability. Various meta-learning techniques are applied to enhance deep recommendation models, especially in scenarios with limited data. The survey categorizes existing methods based on recommendation scenarios, meta-learning techniques, and meta-knowledge representations. It also highlights limitations in current research and suggests future research directions in this area.