Automotive motor bearings are subjected to various complex loads and operating environments, including vibration, impact, and material fatigue, which contribute to a high incidence of bearing failures. This paper presents a self-built cloud-edge collaborative system based on Raspberry Pi and Ethernet as the communication medium, utilizing collected current signals to achieve real-time fault diagnosis for four types of bearing failures: inner race fault, outer race fault, rolling element fault, and cage fault. An integrated motor bearing fault diagnosis system is developed, comprising ‘Sensors—MCU—Raspberry Pi—Cloud Server’, through which data is transmitted from the motor to the cloud server. The received current signals are analyzed through preprocessing techniques such as Empirical Mode Decomposition (EMD), wavelet transform, and adaptive filtering. Subsequently, the processed data is input into a Residual Network (CNN-ResNet) for fault diagnosis. By employing signal inputs, the system can make more accurate fault diagnosis, thereby determining the specific type of real-time fault.