Typically vehicles are generally in movement so number plate detection from vehicles becomes very difficult problem due to possession of images, next to with that number plate may have other designs and some more additional stickers also. within this paper, we have taxing our process for number plate detection which includes many steps, the images of the vehicle’s number plate are composed from the car parking areas and from the network. Firstly, pre-processing of giving input sampled similes are done then feature drawing out of those pre-processed images by applying convolves Gabor and Zonal based algorithm. After that morphological operations are made to find the numbers and also to the segmentation of images, every single character is segmented parallel processing histogram. All the features of every number are designed using zonal based approach. The features are used as put in to the neural networks for quality identification.