Aiming at the technical challenge of accurately detecting ship targets in complex battlefield environments, combined with the characteristics of information complementarity between ship targets and background and interferences in different spectral bands, a multispectral infrared ship detection algorithm based on YOLOv8-n is proposed in this paper. Firstly, the backbone network is designed as a multi-stream network to extract multispectral image features separately, setting the stage for subsequent feature fusion. Secondly, the cross-fertilization residual block (SaCF) is constructed in the feature extraction stage, and the single-spectral segment feature maps are enhanced by capturing the potential correlation between different spectral segments through the self-attention mechanism of the Transformer. Finally, an adaptive hierarchical fusion module (AFM) based on the attention mechanism is designed, which guides multispectral feature fusion by generating fusion weights through the attention mechanism to improve the network's ability to recognize ship targets. The experimental results show that compared with the single-spectrum ship target detection algorithm, the detection accuracy of the multispectral fusion ship target detection algorithm proposed in this paper can reach 94.8%, which is 6.1% higher. The improved algorithm exhibits superior detection performance and is capable of handling ship target detection tasks in complex environments.