The aim of this work was to contribute to the improvement of a methodology for the evaluation of the visual perceptibility of sink marks on surfaces of injection molded parts by machine vision. Based on previous research work on the detection of surface defects a new methodology for the evaluation of the visibility of sink marks using digital images was developed. The methodology uses a surface model to approximate the intensity matrix of captured images. Calculated surface fit functions were used to determine the amplitudes of the second derivatives (ASD) for quantifying the perceptibility of the sink marks. Injection molding parts produced by using different predefined processing parameters were inspected and the influence of given parameters on the visual perceptibility of the sink marks was evaluated. ASD emerged to be an appropriate parameter for the application in a machine vision system. The results partly contradict findings of other authors.