To address the License Plates (LP) extraction in complex natural environments，a novel approach to LP extraction was presented，which utilized LP’s color information，gray feature and geometry characteristics.Firstly，a color classification method was proposed based on the distance between different colors according to the Chinese car license plate color features.Secondly，LP’s background color zones were segmented by scanning lines of picture and analyzing the line segments.Finally，these Zones were translated into binary image in which LP’s background color was dark and LP’s foreground color was white，and validated by the LP’s gray features.In a certain degree，the method could adapt to LP’s type，size，number and direction and wasn’t limited to the location of car the and the background in the picture.It was tested using the natural pictures，and was approved to be effective and adaptive.