This review summarizes the current state of deep-learning models for protein-ligand binding affinity prediction (PLBAP). The paper discusses various deep-learning approaches, focusing on molecular representations, learning architectures, and model interpretability. It categorizes deep-learning PLBAP models into four types: T_ACNN, T_IMC-CNN, T_Grid-CNN, and T_Graph-GCN. Each model type is described in terms of its features, learning process, and interpretability. The paper also evaluates the performance of these models on benchmark datasets, such as PDBbind and DUD-E, and discusses their strengths and weaknesses in terms of scoring and screening capabilities. The review highlights that while deep-learning models have shown great potential in PLBAP, they still face challenges in interpretability and computational efficiency. The paper concludes that further research is needed to improve the performance and interpretability of these models, particularly in the context of drug discovery.This review summarizes the current state of deep-learning models for protein-ligand binding affinity prediction (PLBAP). The paper discusses various deep-learning approaches, focusing on molecular representations, learning architectures, and model interpretability. It categorizes deep-learning PLBAP models into four types: T_ACNN, T_IMC-CNN, T_Grid-CNN, and T_Graph-GCN. Each model type is described in terms of its features, learning process, and interpretability. The paper also evaluates the performance of these models on benchmark datasets, such as PDBbind and DUD-E, and discusses their strengths and weaknesses in terms of scoring and screening capabilities. The review highlights that while deep-learning models have shown great potential in PLBAP, they still face challenges in interpretability and computational efficiency. The paper concludes that further research is needed to improve the performance and interpretability of these models, particularly in the context of drug discovery.