剩余寿命预测是设备故障预测与健康管理的重要组成部分，准确预测设备剩余寿命将有助于决策者及时制定维修策略，避免巨额的失效损失。随着物联网和大数据技术的发展，实时监控设备的运行状态变得越来越便利，为改进和发展设备剩余寿命预测方法奠定了坚实基础。基于相似性的设备剩余寿命预测方法不需要假设设备退化模型，而是通过大数据分析设备间运行状态的相似性进行预测，具有很高的预测精度和强鲁棒性。本文从健康指标构建、相似性度量选取、数据融合等方面，对基于相似性的设备剩余寿命预测方法进行了梳理总结，分析比较了相应性质。文章最后对基于相似性的设备剩余寿命预测方法的未来研究方向进行了探讨。

Remaining useful life prediction is an important part of equipment prognostics and health management. Accurately predicting the remaining useful life of equipment will help decision makers make timely maintenance strategies and avoid huge failure losses. With the development of the Internet of Things and big data, it becomes more and more convenient to monitor the operating status of equipment in real time, laying a solid foundation for the development of remaining useful life prediction techniques. The similarity-based remaining useful life prediction approaches do not need to assume the degradation model, but predict remaining useful life by analyzing the similarity of operating states between the equipment through big data, which have the characteristics of strong robustness and high precision. In this paper, the similarity-based remaining useful life prediction approaches are summarized from the aspects of health index construction, similarity measure selection, data fusion, etc., and the corresponding properties are analyzed and compared. Finally, the future research directions of the similarity-based remaining useful life prediction approaches are discussed.