Increasing use of wireless sensor networks (WSNs) provide opportunity for developing low cost smart systems for structural health monitoring such as bridges. Using WSN, a large number of sensors can be deployed for damage detection. In literature we find many simulation studies to study the parameters for bridge monitoring and damage detection. There are very few works that are deployed in real world scenarios. In this paper a set of commercially available Mica2 motes equipped with accelerometer are deployed on two bridge models in the laboratory to collect the vibration response from the bridge models. The vibration response data collected on intact model is compared with the damaged model. The thresholds are established using the histogram approach on the collected vibration data for intact bridge model. The results show the proper characterization of damage detection.