In this work, we propose an innovative adaptive recommendation mechanism for smart parking. The cognitive radio (CR) module will transmit the vehicle location information and parking space requirement to parking congestion computing center (PCCC) when the driver need to find a parking space. Moreover we use cellular automata (CA) model mechanism for those parking spaces in the parking lot, it can adjust the situation that some parking lots are full and some are not. Here the PCCC can compute with vehicle location information, nearest parking lot, parking lot status and the same or opposite driving direction. We take the driving direction into consideration can reduce road congestion and speed up finding a parking space from vehicles turning around. The recommendation will be sent to drivers after computation and analysis by PCCC and through the wireless communication CR model. The current study evaluates the performance of the approach by conducting computer simulations. Simulation results the strengths of the proposed smart parking mechanism in terms of increased congestion avoiding and decrease parking space finding time.