We here proposed a system where symbol is recognized by the system, user will input symbol images system will apply algorithm to identify the symbol. Here in this system, we applied some image processing steps in order to work with images. We converted the RGB image into gray scale image. Image is converted into gray scale image to apply further image processing steps. Then gray scale image into black and white image, this is done due to accuracy. We had applied some image preprocessing steps in order to remove some unwanted objects and environmental interference. Here we store symbol templates in one directory each image will be of fixed size so that it would be easier to recognize correct symbol. These template images will be in black and white form. System will create dataset of these templates. User will provide query image system will resize the query image. System will finally compare the query image values and template image values in dataset and will display the result in text format. System will take image as an input and output the result in text format. In order to recognize the symbol, we had used the concept of Optical Character Recognition. System will provide result with 60%-80% accuracy.