Rapid, accurate, and nearly real-time aftershock forecasting has attracted increasing public and social attention in dealing with disaster risk and taking effective disposal measures after the mainshock. Many aftershock forecasting methods are seriously affected by catalogue incompleteness in the early stage after the mainshock, which makes it difficult to carry out aftershock forecasting with a disaster reduction effect in time. In recent years, with the development of technology and models, the forecasting of early aftershocks has become possible. In this study, aiming at the "bottleneck period" of aftershock forecasting in the early stage after the mainshock, we elaborated the matched filtering technology and deep learning technology from the perspective of improving aftershock detection rate, the bi-scale empirical transformation technology from the perspective of statistical seismology, and the research progress of the Omi and Lippiello models from the perspective of maximizing the use of aftershock information for real-time forecasting. We analyzed the advantages and disadvantages of various methods and proposed a technical route to comprehensively solve the "bottleneck period" of aftershock forecasting in the early stage after the mainshock. This study provides a scientific reference for researchers to engage in microearthquake detection, aftershock forecasting, and post-earthquake trend research.