Aming at the problem that the detection accuracy of infrared small target is not high in changeable and complex background environment, the advantages and disadvantages of typical algorithms are analyzed, the characteristics of infrared small target images are studied, and an infrared small target detection algorithm based on feature enhancement and multi-scale feature fusion is designed. A feature enhancement module based on multi-scale feature fusion is proposed to enhance the expression of target features. A multi-scale feature information fusion module enhances the target boundary and theme information. A feature fusion decoding module realizes scene and target segmentation. Experimental results show that the infrared small target detection algorithm based on feature enhancement and multi-feature fusion can segment infrared small targets in complex background. In IRSTD-1K dataset, the intersection over union is 60.25%, false alarm rate is 19.6×10^-6 and probability of detection is 85.71%. In NUDT-SIRST dataset, the intersection over union is 87.13%, false alarm rate is 9.14×10^-6 and probability of detection is 96.08%. Compared with other existing methods, it has certain advantages and strong segmentation performance.