Water spot detection in fabrics is conventionally reliant on direct visual assessment, a practice ingrained in our daily lives and textile inspection field. While valuable, this method’s accuracy can be significantly affected by factors like fabric color and patterns resulting from additive textile dyes. In response, we present a detection method employing near-infrared (NIR) hyperspectral imaging, effectively reducing the influence of textile dyes and highlighting water spots. Initially, we acquired NIR hyperspectral images of fabric samples post-dyeing and moisture content conditioning. Remarkably, we observed that typical textile dyes minimally affected the NIR reflectance within the wavelength range of 1265–1626 nm, with a discernible decrease correlated to higher fabric moisture content. Subsequently, we captured NIR hyperspectral images of nine fabrics exhibiting varied colors and water spot characteristics. These images were then transformed into grayscale representations. Among them, the NIR images at 1459 nm emerged as preferred feature images for water spot detection, determined through analyses of contrast, entropy, and principal component images. While conducting a spray-rating evaluation of water spots in water-sprayed samples, comparing them with their respective feature images; the results affirm the efficacy of our developed feature images in facilitating visual identification and analysis of water spots in diverse fabrics, ultimately enhancing the accuracy of fabric water-repellency evaluation.