With the large -scale application of electrochemical lithium battery energy storage storage storage stations and mobile energy storage vehicles, the safety of lithium batteries has attracted increasing attention. Because the lithium battery is very short from thermal abuse to the fire explosion time, how to perform real -time monitoring of the thermal state of the battery in such a short period of time and the early warning related to the energy storage station. In order to strengthen the safety of the lithium battery energy storage system, this article proposes an early early warning technology of lithium battery-based lithium battery-based types of lithium battery models based on BP-SNN fusion neural network models, that is, through multi-sensor fusion technology such as sound waves, using the edge to deploy BP-SNN-based neural network models to shorten the communication delay of sensor data to the computing center, improve data processing speed and efficiency, thereby greatly The early warning accuracy of lithium batteries was improved, and the time of early warning provided valuable time for subsequent fire protection. The results of the simulation results and the actual design of the device show that the method proposed in this article can shorten the warning time by 1.5 seconds, and the accuracy of the early warning to 99.2%can be increased to 99.2%.