Abstract: Anomaly detection technology has an important engineering practical significance for battery data feature mining, retired battery cascade utilization screening grouping and battery operation state safety evaluation. Therefore, this paper proposes a new method of genetic optimization anomaly detection based on the cluster analysis frame. In this method, the cluster analysis is focused on as a center for anomaly detection and the swarm intelligence optimization algorithm is applied as an effective way to solve the global optimization problem. By reasonably designing objective functions to describe data anomaly, the effective detection of abnormal data is realized. Finally, taking the abnormal state detection of the battery data as an example, by comparing the existing methods and the abnormal detection results under the three clustering ideas proposed in this paper, the advantages of the proposed method in its personality, flexibility and accuracy of abnormal detection are verified, especially showing a better detection effect in the clustering optimization detection process based on density idea, It provides a new idea for real-time battery abnormal state detection and data cleaning.