Machine learning can be learned from the image data to characterize the powerful. If the target detection and tracking technology is combined with the machine learning technology,the performance of rapid target detection,accurate identification and precise tracking can be significantly improved,which meets actual needs. In this paper,the YOLOv4 target detection algorithm based on deep learning is applied to the target detection in this scene； then the KCF target tracking algorithm based on correlation filtering uses correlation filtering methods to transform calculations into the frequency domain,reducing the amount of calculations and improving the real time performance of target tracking,comparing the artificial features with the deep features obtained through deep learning and the classifier trained by deep learning,the accuracy of target tracking will be greatly improved. The correlation filter is trained through a multi layer network,which combines correlation filtering and deep learning to balance the real time and accuracy of target tracking,and is applied to target tracking in this scene. The simulation experiment results show that the air target detection and tracking technology proposed in this paper has a target detection accuracy of 95% and a tracking accuracy of 99%,which can realize real time tracking of air targets.