Abstract: In order to optimize the control of freeway network traffic flow, MATNET model was used for traffic flow modeling, and nonlinear optimal control was applied to build on-ramp coordination control model. Aiming for shortening travel time and reducing the on-ramp queue length, global real-time traffic data was used in this model, which can improve the operating efficiency of network traffic, and deal with incidental congestion effectively. Considering the nonlinearity of the model, genetic algorithm was used to optimize performance indexes to explore the effect of optimizing different performance index on traffic flow parameters change. At last, the surrounding network traffic of Nanjing was taken as an example to indicate the feasibility of the proposed model and optimization method.