The detection of aircraft targets in the airport flight area under complex meteorological conditions has emerged as a challenging issue in the research on airport aircraft target recognition. Aiming at the problems of high difficulty and low accuracy in the detection and recognition of aircraft targets under complex weather scenarios such as rain, snow, sand and dust, heavy fog, and at night, this paper proposes an improved algorithm based on YOLOX-S. Initially, a depthwise separable convolution module is incorporated to reduce the parameter count of the backbone network. Then, a Dilated Spatial Pyramid Pooling (DSPP) module is introduced. By combining dilated convolutions, it extracts multi-scale features from the input, generating semantically rich output features. Next, a Dual-Channel Attention (DATTN) module is added to enhance the algorithm’s focus on aircraft targets’ multi-channel features. Finally, a Feature Enhancement Module (FEM) is integrated to fuse shallow feature details, improving the detection of small aircraft targets. Experimental results show that the new algorithm has increased the overall detection accuracy by 10.8% compared to the original, and outperforms other mainstream algorithms of the same type in terms of both detection accuracy and parameter quantity.