In this study, we propose a solution for building an Intelligent Parking Support System. This system not only has characteristics as other information systems but also acts as an intelligent system which integrates three recognition techniques, including automatic recognition for motorcycle license plate (using Cascade of Boosting with Haar-like features and Support Vector Machines - SVM), barcode recognition, and semiautomatic recognition via surveillance cameras. 

Experimental results show that the proposed system is robust and the models work well in three recognition stages. In the license plate area recognition stage, the model gets 99% of accuracy when using 750 images for training and 243 images for testing. In the letter area recognition stage, the model achieves 95.88% of accuracy when using 11,866 and 4,755 images for training and testing, respectively. In the classication stage, we train the SVM model using 2,603 records and test it using 1,550 records to reach the accuracy of 98.99%. The proposed system has full features and can be deployed in practice.