Smart parking is a common demand of citizen, especially for people living in a smart city. It is an important issue since it not only determines the required parking time of drivers but also impacts the urban population and traffic congestion. In this paper, an intelligent parking algorithm is presented based on the predictions of traffics and drivers’ behaviors. The proposed parking algorithm analyzes the historical parking records, predicts the parking traffics and the driver’s parking length and then schedules the vehicles to the parking grids such that the maximal benefits can be obtained. The proposed algorithm also dynamically allocates their reservations but guarantees the parking reservations for the VIP members. But based on the parking space resource. Performance analysis through extensive simulations demonstrates the efficiency and practicality of the proposed scheme.