The trustless deflection occurred by the acquisition system is the main reason to bring the illusive alarms (above 80%). Although the acquisition system of the bridge structural health monitoring system (BSHMS) is designed by the advanced sensors, for example, the photoelectric imaging deflection meter system and so on. However, the trustless deflection generated by much reason in data acquisition system and transmission system, and these trustless data are unavoidable. A novel method based on the correlation analysis of bridge?s checking points and the RBF neural networks is proposed for correcting nonlinear deflection trustless data. Compared with conventional methods (its MSE is 4.3815), the proposed approach (its MSE is 0.7973) assures more accurate and accords with practice. Simulation results verify the effectiveness of the designed method and theoretical discussions.