This paper introduces an individual-based simulation model to study the land development process in San Diego by using multi-agent simulation (MAS) software, where the urban expansion activities are realized through the construction behaviors of three major types of developer agents (residential/industrial/commercial). Except compiling codes for the interaction activities between different kinds of agents and their surrounding environment, this model also includes the analysis function of urban renewal into the simulation process. And in order to find a suitable land policy for urban growth management in San Diego, this paper conducts a prediction research about future land development in three different types of administration scenarios, and evaluates their simulation layouts from two separate perspectives: visual observation and geometrical calculation. After that, following the suggestions of comprehensive plan in 2050 and smart growth principles, this paper proposes a series of measures for the improvement of current land management, and verifies their effectiveness through the comparison research between optimized scenario and other three administration policies mentioned above. Also, for the purpose of increasing simulation layout’s consistency with the real world, several possible ways are put forward to optimize the modeling procedures at the end of this paper.