Abstract: In 2015, the United States put forward the concept of precision medicine, which changed medical treatment from "one size fits all" to personalization, and paid more attention to personalization and drug customization. In the same year, Spritam^®, the world's first 3D printed tablet, was in the market, marking the emerging pharmaceutical 3D printing technology was recognized by regulatory authorities, and it also provided a new way for drug customization. 3D printing technology has strong interdisciplinary and high flexibility, which puts forward higher requirements for pharmaceutical staffs. With the development of artificial intelligence (AI), modern society can perform various tasks, such as disease diagnosis and robotic surgery, with superhuman speed and intelligence. As a major AI technology, machine learning (ML) has been widely used in many aspects of 3D printing drug, accelerating the research and development, production, and clinical application, and promoting the new process of global personalized medicine and industry 4.0. This paper introduces the basic concepts and main classifications of 3D printing drug, non-AI drug optimization technology and ML. It focuses on the analysis of the research progress of ML in 3D printing drug, and elucidates how AI can empower the intelligent level of 3D printing drug in pre-processing, printing, and post-processing process. It provides a new idea for accelerating the development of 3D printed drug.