This study proposes an intelligent decision support system for guiding available parking slots on the roadsides in urban areas. The front-end system is used to develop an application that assists drivers in parking their cars. The main goal of this study is to solve the problem that the drivers spend a substantial amount of time in finding suitable parking slots in busy city streets. However, the extended problems, including traffic congestion, additional vehicle emissions, and illegal parking, can also be simultaneously solved. Meanwhile, the architecture of the proposed system can reduce the installation and maintenance costs compared with the existing schemes. The architecture of the intelligent parking system includes four stages: (i) the system utilizes sensors to capture the pictures of on-street parking and send them back to the database server in a short period; (ii) the proposed method uses an objection detection algorithm (i.e., region-based convolutional neural network) to identify whether the on-street parking slots are available or not and count the number of available parking slots; (iii) a smart parking application is developed for the driver to obtain the real-time parking information by using any possible mobile devices; (iv) two mathematical models corresponding to two situations are proposed for the recommendation of the best parking slots. The proposed system is also an important basis in a smart city. The numerical experiments confirm that the proposed algorithms embedded in the system are also efficient and effective compared with the current commercial solvers. [Link to this article...]