Identifying essential factors in cellular interactions and organized movement of cells is important in predicting behavioral phenotypes exhibited by many bacterial cells. We chose to study Myxococcus xanthus, a soil bacterium whose individual cell behavior changes while in groups, leading to spontaneous formation of aggregation center during the early stage of fruiting body development. In this paper, we develop a cell-based computational model that solely relies on experimentally determined parameters to investigate minimal elements required to produce the observed social behaviors in M. xanthus. The model verifies previously known essential parameters and identifies one novel parameter, the active turning, which we define as the ability and tendency of a cell to turn to a certain angle without the presence of any obvious external factors. The simulation is able to produce both gliding pattern and spontaneous aggregation center formation as observed in experiments. The model is tested against several known M. xanthus mutants and our modification of parameter values relevant for the individual mutants produces good phenotypic agreements. This outcome indicates the strong predictive potential of our model for the social behaviors of uncharacterized mutants and their expected phenotypes during development.