Any distribution of incentives lead to multiple modes of response, it is thus very difficult to stimulate single mode of vibration in experiments. The limitation of current signal processing techniques implies that traditional methods used to estimate the damping ratio cannot effectively separate the superposition modes, which leads to larger error in obtained damping ratio. Starting with modal superposition method theory in dealing with multiple-degrees-of-freedom system dynamic response, it was pointed out that mode confusion problem is one important limitation in precise damping ratio test. Modes are more crowded and the superimposed effect is more significant when system damping is larger and stiffness is lower. One program was proposed to achieve “pure mode” extraction by applying modal truncation in numeral calculation. The formula expressions of frequency-response spectrum peak line in both of resonance excitation experiment and numeral calculation were derived. The relationship between them was investigated. “Pure mode” calculation results were used to correct experimental damping ratio. Through four different board units, eight pre-order experimental modal analysis and numerical calculation parameter correction, combining frequency response function to verify data reliability of the correction damping ratio, some laws of damping ratio between different structures or materials were obtained. The results show that: the identification accuracy of composite plate’s modal damping ratio is lower than steel’s by model experiment calculation, and its damping performance is often underestimated and large amplitude correction are also often needed. The method provides a guideline for further study of modal parameter identification.