Rolling mill machinery plays a crucial role in industrial production. To minimize equipment downtime and maintenance costs and achieve cost reduction and accident prevention, online fault diagnosis of key equipment is crucial. This paper focuses on the vibration signals of rolling bearings and develops an online monitoring and fault diagnosis system. The system consists of real-time monitoring, fault information management, fault diagnosis module based on time-frequency domain analysis, and fault diagnosis based on deep transfer learning. It enables real-time recording, display, and storage of critical equipment vibration signals and performance indicators, conducts online analysis, and utilizes multiple common bearing databases within the fault information management module for calculating, storing, and querying bearing fault frequencies. The fault diagnosis module integrates traditional time-frequency domain analysis and deep transfer learning methods based on the Resnet-18 network for in-depth fault diagnosis and trend prediction. The experimental results show that the system has a friendly interface and simple operation and can effectively apply time-frequency domain analysis functions and deep transfer learning algorithms to new machines lacking sufficient fault data, achieving high recognition accuracy.