Due to the traditional parameter optimization methods hard to realizing the effective improvement of the overall performance of the bridge structure after optimizing the bridge structure parameters, a research on the optimal design of bridge structure parameters based on multi-level neural networks is carried out. By establishing a finite element model of bridge structure, based on the multi-level neural network, the optimization goal of the bridge structure parameters is determined, the sample data for the optimization of the bridge structure parameters are collected, and the multi-level neural network parameters of the bridge structure are optimized and solved. A new parameter optimization method is proposed. The experiments have further proved that the new parameter optimization method can effectively improve the comprehensive performance of the bridge structure compared with the traditional parameter optimization method, which provides a new optimization idea for increasing the efficiency and reducing the consumption of bridge construction.