Mixed traffic is a main characteristic of Chinese urban traffic. Motorized vehicle, non-motorized vehicle and pedestrian own different specialties and interference among them is seriously. Microcosmic model is an efficient research tool for mixed traffic, but now the research in it is still absent.The dissertation focuses on the study of urban mixed traffic microcosmic modeling and corresponding simulation software system. The main contents and contributions are listed below:1. Establishing a new microcosmic model for motorized vehicle: mutual action model. Aim at the shortcoming of car-following model, thinking over what is the essence of the driver's acceptance of stimulation and reaction again. Analyzing the phenomena that vehicle follows the leader in press and safety. The paper considers that the movement state of vehicle is the sum of the two contradictions one called pressure and the other called safety. Based on the view the dissertation establishes a new microcosmic model for motorized vehicle: mutual action model and discusses its stability. Through math and simulation the dissertation compare the mutual action model with car following model in several aspect. The result is: the two models is same in car following, and the mutual action model can describe acceleration unbalance, different driving decision, stop, start, overtaking of motorized vehicle widely but car following model can't describe these characteristic in an unit style. Then it is concluded that the mutual action model is well than car following model.2. Establishing a new model for non-motorized vehicle: vector field model. Based on the theory of field and considering all of the effects from around movement objects as one object in center of mass, expressing velocity and acceleration as vector, the dissertation extends mutual action model to a new microcosmic model for non-motorized vehicle: vector field model. The simulation shows that the non-motorized vehicle moves in snake track and stop in insert style. It corresponds with the characteristic of non-motorized vehicle. The highest density of traffic flow from line regression of its statistic data is in the area of factual data. Then it is concluded that the vector field model is effective in non-motorized vehicle traffic.3. Realization the interference in mixed traffic. Through analyzing the reason and representation of the interference in mixed traffic, the dissertation represents the methods of realization and confirms its parameters by a factual example.4. Establish a mixed traffic simulation system. Aim at the mixed trafficsimulation is not exist, explore the simulation technique for mixed traffic. The dissertation establishes an Urban Traffic Simulation and Analysis System (UTSAS). This system can describe the common movements of different objects in urban road and can get many statistics from the traffic simulation result.Finally, the dissertation concludes with a summary and perspectives of future research.