A system and method for differentiating pictures and texts includes, first, dividing an image corresponding to an image file to a plurality of unit areas of the same size each containing a plurality of pixels of the same number; generating a threshold by a threshold setting module based on the relationship between a gray-level and quantity; a picture text differentiating module sequentially performing statistics of the gray-levels of the pixels of the unit areas; sequentially comparing the gray-levels with the threshold; totaling the times that the gray-level equals to the threshold; and the picture text differentiating module comparing the times of equality with a preset comparison value to differentiate whether the data corresponding to the unit area is picture data or text image data.