Existing image super-resolution networks are mostly designed for visible light images, with relatively fewer studies focusing on infrared image super-resolution, and most of them simply adopt methods from visible light image super-resolution. In response to the low resolution and blurred edges of infrared images, a gradient-guided infrared image super-resolution reconstruction network was proposed. The gradient information in low-resolution infrared images was fully utilized by the network, fusing the gradient map with the extracted features, thereby resulting in a high-resolution image with clearer edges and higher contrast. The experimental results of the comparative and ablation studies demonstrate that the proposed method outperforms other comparative methods in infrared image super-resolution reconstruction, generating high-resolution images of higher quality.