The fault diagnosis of petrochemical units is of great significance to the reliability and safety of modern industrial systems.Artificial intelligence (AI) technology as an emerging field in industrial applications and an effective solution for fault identification,it has received increasing attention from academia and industry.However,under different operating conditions,the AI approach faces great challenges.The artificial intelligence algorithms in fault diagnosis of petrochemical units were described from two aspects:theoretical background and industrial application.Different artificial intelligence algorithms were introduced,including K -nearest neighbor,naive Bayes,support vector machines,artificial neural networks and deep learning methods.Then,an extensive literature survey of AI algorithms in industrial applications was carried out.Finally,the advantages,limitations and practical enlightenment of different AI algorithms were summarized,indicating the key role of technological progress,multi-modal data integration,real-time monitoring and prediction and algorithm universality in improving the efficiency and reliability of the petrochemical industry,and looking forward to future research directions and challenges.