The visible light small target detection and recognition technology based on deep learning algorithm is an important field of academic research. Small targets account for a small proportion of all targets, have unclear features, and are difficult to detect. Firstly, in order to achieve fast and accurate recognition of small visible light targets, a deformable convolutional network is used based on the YOLOv8 network model to adjust the shape and size of the receptive field adaptively, enhancing the network’s ability to learn the invariance of complex targets. Secondly, an attention module is added to enable the model to focus on important information and improve its robustness. Finally, a specific small target dataset is constructed for training and validation, and the results show that the improved model has high accuracy and recall for specific targets.