Abstract: Based on the vibration caused by moving vehicles,the pattern recognition method was adopted to identify the damage detection,localization and severity estimation for bridge structures.With the technique,the dynamic responses of a bridge under the same moving vehicle for various predefined damage locations and severities are figured out,and the responses at testing points are written down to form a damage pattern database.Then the measured responses under moving vehicles,collected from the instrumented sensors on a bridge structure,are compared with the damage patterns,and the nearest method for clustering analysis is engaged to carry out the damage detection,localization and severity estimation.The comparison is displayed in a contour or image plot with damage location and severity as independent coordinates.By taking the damage detection of a simply supported beam under a constant moving load with a constant speed across the structure as an example,the concept,effectiveness and robustness of the pattern recognition were demonstrated numerically.The numerical results show that the proposed technique is effective and robust to estimate the damage location and severity of bridge structures.