This paper presents artificial intelligence techniques for multi-spatiotemporal urban growth modeling. The objective of the study is to simulate and predict urban dynamic changes and quantify their impact on different land uses. The design of the algorithm covers three stages: crisp cellular automata modeling, genetics algorithm calibration, and fuzzy logic module. Crisp cellular automata define a set of transition rules as a function of spatial neighborhood structure and input data. The rules serve as a modeling engine to simulate spatiotemporal urban dynamic changes. The calibration of such rules is performed spatially on a township basis and temporally as a function of time. Three measures are developed as part of an evaluation scheme for model calibration. The second stage is using genetic algorithm to increase the efficiency of the calibration process. Parameters for genetic algorithm, including string design, encoding, selection criteria, crossover and mutation are discussed. The final stage introduces fuzzy logic to provide good initials for crisp model and to model the semantic knowledge/concept. Fuzzy inference preserves the continuity of urban dynamics spatially. Fuzzy parameters namely, fuzzy membership functions, fuzzy rules and fuzzification-defuzzification process are designed. The model is tested over Indianapolis city, IN for a test period of 30 years. Input data include digital elevation model, roads, land cover and population density in raster format. Modeling results indicate good accuracy with close urban match patterns between real and simulated data. The model can be used to quantify both the magnitude of urban growth around installations and the effect of urban growth on critical installation mission activities. Our model makes use of the historic multitemporal data in land use to help the decision makers build their decisions based on an optimal real time results.