Aiming at the detection problem of point like maneuvering targets in infrared sequential images,an infrared dim and small targets detection method based on SSD framework is proposed,which can effectively perceive the difference between small and small targets and interference by taking advantage of the convolutional feature,which has a more detailed description ability to the images.In this paper,the SSD detection framework for dim small targets is designed,and it overcomes the difficulty that traditional methods lack a large amount of experimental data.Simulation verification is carried out by using real infrared image database.Experimental results show that this method has good detection effect and false alarm suppression ability for targets with low SNR in various scenes.