Abstract: With the expanding of big data and artificial intelligence (AI) technology in oil and gas exploration and exploitation, the development of digital and intelligent lost circulation control technology has become an inevitable trend, and the core lies in the machine learning-based algorithm model and the bundled software. This paper systematically summarizes and analyzes the application status of artificial neural network (ANN), support vector machine (SVM), random forest, case-based reasoning (CBR) and other machine algorithms inlost circulation feature prediction, real-time monitoring of lost circulation and application decision-making model. Additionally, a comparison is made on the input parameters, output parameters, test accuracy, and application effects of various machine learning algorithms. On this basis, it has been found that machine learning algorithms tend to have good application prospects in thief zone location prediction, lost circulation monitoring and early warning, and recommendation of measures for preventing and plugging lost circulation. Compared with artificial statistical analysis, the machine learning algorithms have obvious advantages in timeliness, accuracy and large-scale application. However, it is still unable for us to conduct scientific prediction and calculation of the loss pressure, pore/fracture size and other key parameters of lost circulation features and to optimize the field construction using these algorithms. As we know, foreign oil and gas companies made an early start for the overall arrangement of digital drilling and completion technology. At present, they have developed the software for lost circulation prevention and plugging by integrating a variety of machine learning algorithms, and have achieved certain application results in the field. However, China started late in the management of data related to lost circulation prevention and plugging, the development of machine learning algorithms and the research of bundled software, and it has not yet established any mature, reliable digital platform and intelligent expert system for lost circulation prevention and plugging. In order to accelerate the digital and intelligent transformation development of lost circulation technology in China, the research should focus onthe following three aspects:(1) to promote the integration of multi-dimensional data related to lost circulation, build a data lake involving earthquake, logging, drilling, laboratory evaluation and field construction of lost circulation, and make up and improvethe shortcomings of data; (2) to strengthen the explanatory study on machine learning algorithm model, and improve the scientificity and accuracy of the algorithm model based on the mechanisms of lost circulation; (3) to integrate the lost circulation data lake and algorithm modules, establish an intelligent expert system for the intelligent prediction and early warning as well as assistant decision-making of lost circulation in different regions, and formulate fine operation standards for lost circulation prevention and plugging, so as to comprehensively improve the one-time success rate of fd lost circulation control.