Abstract: This paper proposes a big data-based anomaly analysis method for electricity inspection, integrating random forest, SVM, and time series analysis. Experimental results show it outperforms traditional methods in accuracy and response time.