Abstract: Through model testing, damage identification techniques for suspension bridges are investigated. A 10m long suspension bridge model was designed and manufactured for the purpose of damage identification testing. Using the physical model for error analysis, a corresponding high accuracy finite element model was developed for the test model. The damage identification strategy is determined using the parameters provided by the monitoring system and/or tests on the suspension bridge, and considering the sensitivity and correlation of these parameters with damages. The FEM was used to develop possible damage cases which provided the training sample data for the BP network, and then the testing data were physically modeled on the test model. For the damage cases simulated, damage locations were identified correctly 86% of the time, with acceptable accuracy in identifying the degree of damage. The method has good prospects for application. It also has significant implications for health status evaluation of suspension bridges using structural monitoring.