With the advancement of infrared detection technology and unmanned reconnaissance platforms, the capability of long-range target detection and recognition has been significantly enhanced, posing new challenges to traditional camouflage techniques and evaluation systems. To achieve quantitative evaluation of camouflage performance, it is necessary to consider multi-band information and incorporate environmental factors from practical applications. This paper proposes an infrared camouflage assessment method that integrates visual saliency and texture features. The method employs a graph-based visual saliency (GBVS) model to extract saliency features of the target and background and incorporates texture features of the target region to construct a unified evaluation metric through linear weighting. Experiments based on multi-band infrared images (short-wave, mid-wave, and long-wave) are conducted under different camouflage states, observation angles, and temporal conditions. Results demonstrate that the proposed method exhibits good stability and discriminative capability across various imaging conditions, and the evaluation outcomes are highly consistent with human visual perception. This study provides theoretical and engineering support for the development of infrared camouflage materials and the optimization of infrared target detection systems under multi-band conditions.