When using traditional methods for detecting overheating faults in substation operation equipment, environmental interference often leads to a decrease in the accuracy of overheating warnings. To solve this problem, an innovative method based on multi-scale feature fusion is proposed in the paper for detecting overheating faults in substation operating equipment. This method uses improved KPCA to perform dimensionality reduction on the characteristics of substation operating equipment; Utilizing multi-scale feature fusion to detect overheating faults in substation equipment. The experimental results show that this method can provide effective warnings and achieve reliable detection results under any time conditions, and has great application value.