Abstract: With its significant breakthroughs in recent years, deep learning technology has been used in medical, pharmaceutical and many other areas. This review focuses on the development and application of deep learning for innovative drug discovery, summarizes typical cases of deep learning for the prediction of protein structure, drug target and drug-target interaction, the design of drug synthesis route, de novo drug design, and the prediction of drug absorption, distribution, metabolism, excretion and toxicity (ADMET), and discusses the current problems and possible solutions, in the hope of providing some reference for the development and application of deep learning for drug discovery.