A method for adjusting the color saturation during electronic image processing, in which a two-dimensional original image is electro-optically scanned along rows and columns for three primary colors and the resulting image signals are transformed into a luminance signal Y and two color-difference or chrominance signals C1, C2. A pre-selected adjustment of color saturation is used, in which, in accordance with a non-linear characteristic curve, small chrominance signals C1, C2 are more amplified than larger chrominance signals C1, C2 corresponding to a higher degree of saturation.