In Vietnam, current popular parking management systems only collect and display license plate data, leaving humans to verify the driver's identity. Manual authentication has several disadvantages, including longer processing times, congestion, unexpected mistakes when human health is poor, and high labor costs as well. In this study, we provided a solution to the above problem when deploying an automatic motorcycle keeping system that uses artificial neural networks to process and recognize facial images and license plates. The accuracy of the system is evaluated using data collected from the real motorbike keeping lot. The system has high accuracy and fast processing speed when it is able to recognize 93% of license plates, and the rate of identification and matching of the sender and receiver's faces is as high as 84%. The proposed system outperforms current technologies in terms of reliability and safety.