The wrinkled fabrics will miss their wrinkle properties with finishing processes by time assuming and their wave will be reduced. The goal of this study is finding fabrics wrinkles without missing that during usage. In these types of fabrics the wrinkles are not decreased after washing and ironing and after heat finishing due to fabric initial properties; it will be back to primary condition. In this research the surface fabric waves were analyzed and wave's prosperities were extracting, by using wavelet analyzing and then these results were used to prepare the wrinkled fabric. So it comes down to study is finding the specific defects on the fabrics. Two (100%) cotton and the intermingled hybrid yarn were used for this matter. After surface imaging, the fabrics were analyzed by wavelet and features were determined. Then the data was examined by improved artificial neural network (ANN).This net was a suitable method with low error and high prediction ratio and with respect to accuracy of about 90% presents a good method to recognize the fabric surface wrinkles. Results show that picking method to form a desired figure can be determined by training patterns to the net.
