The measurement system with the laser displacement sensor to measure the three dimensional shape of the fabric wrinkle was developed for the objective evaluation of fabric wrinkle recovery. The shape of fabric wrinkle was characterized by amplitude of power spectra and distribution of polar angle through the process of two dimensional fast Fourier transformation and spline. Artificial neural networks designed by five types of the input patterns were used to evaluate the fabric wrinkle objectively. Wrinkle recovery rank evaluated by neural networks showed good relation to rank generated by human experts. Also the wrinkle rank measured by laser sensor system was not much affected by fabric color.