Cellular Automata is a discrete model studied in computability theory, mathematics, physics, complexity science, theoretical biology and microstructure modeling. This paper proposes a model to balance the flux and safety, to maximize the flux on each road whilst ensuring work well with safety based on the Cellular Automata. In order to analyze the flow of traffic, the traffic of three-lane freeway and four-lane freeway are simulated based on the cellular automata model with the factors of velocity and density in the first step. For safety, a rear-end collision model and overtaking lateral collision model were employed to analyze the relation among safety. Reaction time and distance are introduced in the model, and different vehicles are considered. Safety-related factors are also chosen based on back propagation neural network. At last, this paper establishes a model to examine the performance of the traffic flow under the multi-lane freeway in light and heavy traffic. It simulates the situation in real life based on actual data. The optimal relation and optimal value has worked out between flux and safety through simulating the situation in real life based on actual data. To achieve a better model, further discussion, is made on how to adapt to different situations with the keep-left, the unrestricted, and the intelligent system controlled. The experimental result shows that the number of lanes is not an influential factor under any circumstances. The maximum speed limit plays a significant role in light traffic while it is of no importance in heavy traffic. However, the minimum speed limit plays a significant role in heavy traffic while it is of no importance in light traffic.