A system to detect fabric damage introduced by household washing machine was developed by using computer image processing and analysis. The images of Marquisette fabrics used for evaluating the gentleness of laundry machine action in the AHAM test method, were captured in gray levels using an image scanner with resolution of 600 dpi. The images were filtered using Gaussian filter and converted into binary images. The converted images were further treated by thinning method. From the images, the dislocations of cross-over points of the weft and warp were detected. Then fabric damage was evaluated by the ratio of total number of the dislocated cross-over point to the total number of cross-over points in observed area. In addition, the degree of fabric damage was also evaluated by the summation of the gap between the dislocated weft and imaginary reference weft. The results obtained with this image analysis system showed good correlation with those of human observers.