Background <p>Multi-view IFD is crucial in industrial automation. It leverages industrial multi-view data, combined with machine learning and sensor technologies, to automatically identify abnormal conditions in mechanical equipment. This enhances diagnostic efficiency, reduces costs, and ultimately enables proactive prevention.</p> Purpose <p>Given the widespread application of multi-view and multi-sensor processing techniques in areas such as intelligent factories, autonomous driving, robotics, and unmanned aerial vehicles (UAVs), an increasing number of scholars have begun to delve into multi-view information fusion detection (IFD). Therefore, conducting in-depth research on multi-view information fusion detection is imperative. This will provide a reference for researchers and practical guidance for industry practitioners implementing multi-view information fusion detection.</p> Method <p>The article thoroughly analyzes the latest advancements in single-view and multi-view IFD. Traditional approaches to feature engineering emphasize single-view feature selection and feature extraction, while multi-view methods necessitate additional attention to view fusion.</p> Conclusion <p>Considering the influence of different view fusion strategies on fault diagnosis and drawing inspiration from multi-source data fusion, this paper reviews the multi-view fusion from data-level, feature-level, and decision-level fusion. When it comes to fault diagnosis, we categorize fault diagnosis methods into supervised learning and unsupervised learning de- pending on the availability of fault tag information in practice. Specifically, we outline fault diagnosis methods tailored to these two scenarios from the aspects of sensor, machinery equipment and industrial system faults. Finally, the paper discusses the guiding principles and potential challenges in multiview IFD.</p>