Abstract: In response to the complex, multi-source, and precision-oriented fault diagnosis requirements of large-scale industrial processes, data-driven fault diagnosis methods have attracted increasing attention. These methods integrate historical data, real-time data, and multi-source information to significantly enhance the accuracy and efficiency of fault identification and localization. This is crucial for ensuring the safe, reliable, and real-time maintenance of industrial production processes. Firstly, a comprehensive overview on data-driven fault diagnosis methods is given, focusing particularly on steady-state systems, dynamic systems, and industrial processes characterized by nonlinearity and non-Gaussianity. Secondly, distributed fault diagnosis methods for large-scale industrial production processes are reviewed. Based on the distributed structure and sensor networks for large-scale systems, the latest works are focused on system decomposition and data fusion, correlation analysis, and consistency mechanisms. With the aid of the distributed fault diagnosis methods, the monitoring capacities are distributed among all the subsystems, enabling each subsystem to independently assess its safety and performance based on its own data and interactions with neighboring subsystems. This approach offers distinct advantages in monitoring and diagnosing large-scale industrial processes. Finally, the paper concludes by summarizing practical applications evaluating the performance of data-driven fault diagnosis methods. Potential trends in this field are also highlighted, including the integration of qualitative and quantitative methods, the enhancement of diagnosis robustness, and the assurance of data security.