In this paper, a damage detection method based on a combination of wavelet analysis and an interval type-2 fuzzy logic system (IT-2FLS) is proposed. Firstly, synthesizing IT-2FLSs as a data-driven model is proposed. The structure is then divided into elements and excited to be vibrated to measure vibration signal. Average quantity signal of wavelet transform coefficient (AQWTC) of vibration signal with a used-scale-sheet is established. The IT-2FLS is utilized to identify the structure at its undamaged time via AQWTC signal. At each surveying time, AQWTC at each element is calculated to estimate difference of corresponding AQWTCs between two cases: undamaged status and the status at the checked time. By applying the AQWTC’s contrast at two these times, a damage coefficient is described which is used to estimate status of the structure. Besides, in order to predict structure’s status, the time-series prediction using the IT-2FLS and the calculated damage coefficient are also presented. The effectiveness of the proposed method is demonstrated by experiment via data sources measured from dynamic response of a real structure.