In view of the problems that the rotating machinery running state cannot be checked anytime and anywhere in monitoring,the data generated by monitoring is gradually increasing,and the fault features extraction is hard,a cloud platform-based monitoring system for rotating machinery bearings was proposed,taking bearings as the key component.Firstly,temperature and acceleration sensors were used in the system,and STM32 MCU was used to obtain the data needed for bearing monitoring.The data were then transferred remotely using the narrowband internet of things and stored in a cloud database.In the cloud platform,the bearing status was monitored by correlation time-domain and frequent-domain analysis,and a multi-scale attention convolution neural network diagnosis algorithm designed was used for bearing fault diagnosis.Finally,it is indicated that the fault diagnosis algorithm has a high diagnosis accuracy and the system can run well.