Due to the rapid development of information technology both in hardware and software, machine learning has been a very powerful and thus popular tool in dealing with problems in many fields. Based on the data-driven way, complicated mathematical models are not required to derive in employing machine learning techniques for solving problems. Instead, only input-output data are necessary to collect for training machine learning models. Thus, this study aims to collect and analyze literature on using machine learning in integrated circuit substrate electrical tests. Then, cases of applying machine learning in integrated circuit substrate electrical tests are introduced. Finally, potential directions of future study are provided.