With the development and application of intelligent transportation systems (ITS), parking guidance and information system (PGIS) has become a hot topic for research and applications. Parking variable message sign (parking VMS) is the most common form of PGIS currently in use, particularly in providing the space availability information to en-route drivers. The aim of this study is to model drivers' responses to the parking VMSs under various parking information in a dynamical way. To ensure the guiding efficiency and assist drivers to identify parking lots with vacant spaces, the display strategies of parking VMS were studied. A discriminated release strategy with a space occupancy indicator (SOI) was then proposed to determine the display content on the parking VMS, in which an agent-based simulation was introduced to incorporate both parking choice model and traffic assignment algorithm dynamically. Guiding reliability and display strategies were further analyzed to attain a discriminated release strategy with SOI effectively. The results indicate that the guiding reliability was improved by reducing the SOI value of the corresponding parking lot, and the relatively uniform guiding reliabilities of all parking lots were attained by modifying SOI values iteratively. Additionally, the performance by incorporating the waiting time into the parking choice model was investigated and recommended. The discriminated release strategy presented in this study will assist in designing and operating urban PGIS.