Abstarct: Before the smart phones being so ever-present, license plate recognition systems were usually baxsed on a high resolution camera, a powerful computer and a separated transmitter for sending information. In recent years, some of the organizations such as parking management and traffic monitoring companies suggested the implementation of ALPR on Android system in order to offering services by smart phones. In this thesis, we tried to answer to this demand and implement license plate recognition on smart phones. Even though, the speed is not very important in this usage, we tried to offer a quick algorithm for detection and recognition of license plates. To reduce the computational complexity, we used a simple and effective way. First, we use Vertical edge detection and Horizontal projection to distinguish candidate’s areas of license plate. Then characters are isolated with a connected component labeling algorithm and finally a neural network is used to recognize characters of the plate. Furthermore, we used parallel processing to increase the running’s speed of the application. The proposed method was tested on pictures taken from smart phones. At first, this method was tested by C++ language and the OpenCV library in Visual Studio. Then, we implemented it by Java and OpenCV in Android Studio for smart phones. The average of processing time and the percentage of recognition were 0.67 and 96.1% respectively.