Abstract: The application of combined models for traffic demand forecasting is analyzed, and emphasis is put on two combined models applied to demand forecasting for urban traffic. An example of combined model for trip distribution and assignment based on the principle of entropy maximization is presented. A model of generalized trip cost for all types of urban passenger traffic modes is proposed. Then a combined model for trip mode categorizing and assignment is put forward based on the principle of generalized trip cost and dynamic probability assignment.