Traditional electricity safety protection is mainly achieved through safety protection devices, manual troubleshooting, and traditional monitoring and alarm systems. The problem of missed and false alarms is relatively serious. In response to this situation, the author proposes a real-time monitoring and hazard warning system for electricity safety based on the Internet of Things and big data analysis technology. The author first combines harmonic-based load analysis, inverse time characteristics to implement an adaptive warning threshold setting method, and data processing technology based on “data recording”, effectively optimizing the hidden danger warning algorithm and improving the system’s online monitoring and comprehensive analysis capabilities. Secondly, in order to verify the performance of the real-time monitoring and hazard warning system for electricity safety, two methods were used for testing: building a test platform in the laboratory and selecting real-scene test points. Finally, the test results indicate that the system can quickly analyze electrical safety hazards and implement early warnings based on electrical characteristic parameters. In the load analysis experiment on the electricity consumption of student dormitories, the accuracy of load identification reached 86%, which can be put into practical application.