In the context of materials genetic engineering, data-driven machine learning technology is driving materials research into a new paradigm after theory, experiment, and computation, which is the fourth paradigm. Machine learning can make full use of the existing experimental data and can be able to fully exploit the hidden connections behind the data for purpose of achieving a more accurate prediction of material service performance without clarifying the principles behind it. Therefore, machine learning can greatly reduce the time and cost required for experiments, showing amazing vitality in the field of predicting material performance. The service behavior of a material reflects its performance and is one of the key factors affecting its use. Relying on the material base data from previous experiments and the established database, the prediction of the service performance of materials has been initially achieved. Since different machine learning algorithms have a great impact on the prediction accuracy and generalization ability of the results, it is of crucial importance to select a suitable machine learning algorithm. In this paper, we summarize and analyze four common models for predicting the service performance of metallic materials, including Artificial Neural Networks, Random Forests, Support Vector Machines, and Cluster analysis. The development history, advantages, and disadvantages of these four models are briefly described. These four models have obvious advantages in predicting the service performance of metallic materials and designing new high-performance metallic materials. After that, we present the practical applications of machine learning algorithms in predicting several typical service performance of metallic materials. In the field of materials research, in order to save time and cost, the chemical composition of metallic materials, test environment conditions, and other factors can be used as features and input into machine learning training models, which can make accurate and effective predictions of the service behavior of metallic materials, and also provide reliable ideas for designing high-performance metallic materials. This paper introduces the use of machine learning in the three more typical service properties, fatigue, creep, and corrosion, whose testing generally has a long time cycle and high cost. Then we analyze some specific cases, as well as briefly introduce the application of machine learning in predicting other service properties, such as hydrogen embrittlement and irradiation damage. Finally, the characteristics of machine learning for predicting the service performance of metallic materials are summarized, some current problems to be solved are analyzed, and the prospects for its development are presented.