In this simulation, based on a spatial garbage-can model, I treat the urban dynamic system as a set of random-walk elements interacting with each other in time. Five elements are considered: decisionmakers, choice opportunities, problems, solutions, and locations. Decisionmakers, problems, and solutions meet in an unpredictable way, are thrown into garbage cans as choice opportunities in certain locations, and something happens. In contrast to most traditional spatial simulation techniques, in which space is global and static, I consider in this simulation locations of facilities as dynamically floating on the stream of opportunities, interacting with other elements. The simulation results imply that the relationship between problems and choice opportunities dominates the outcome of the evolution, whereas the effects of the spatial structure as well as the structures of decision and solution are insignificant. The model may provide useful insights into how urban dynamics evolve.