Aiming at the drawbacks of poor contrast, fuzzy edge and insufficient texture information widely existing in infrared images, a detail enhancement method of infrared images based on second-generation wavelet cumulative probability is proposed. Firstly, the algorithm decomposes infrared images for the second-generation wavelet mono-layer, and the cumulative probability distribution function is used to calculate the detail histogram equalization of the infrared images, then the histogram equalization is processed respectively. The comparison of several algorithms shows that the proposed method is outstanding in image detail in the digital detail enhancement of infrared image, and the noise suppression effect is good.