A model is developed to determine the optimal mix of traffic types between zones in a large city, provided the distribution of person-trips by purpose is given. The utilization ratio of each of the following four traffic types are determined: train, bus, taxi and private car. The model is a linear-programming model, where the generalized cost obtained as a sum of the trip cost and the travel time converted into a money term is minimized under various constraints, such as the capacity of each transportation facility and the permissible amount of pollutant emission. The outcomes of such drastic policies as the reduction of the total amount of car traffic or the reinforcement of pollution control can thus be easily simulated. By applying to the data of the OD survey in 1970 in Osaka, it is confirmed that the model can explain fairly well the actual situations of modal split among different traffic means. The impact of highly enforced environmental protection control is also evaluated by finding the optimum solution to the model.