Abstract: A potential risk of ultrasonic guided wave testing must be noticed,which the weak echo signal from the small defect may be submerged in the noise signal and cause missed detection.By analyzing the bifurcation characteristics of the Duffing system with the amplitude of the driving force,the critical state of the system periodic jumping can be obtained.The ultrasonic guided wave signal with same frequency as the driving force is added into Duffing system as the disturbance term of the driving force.This is equivalent to changing the amplitude of the driving force,which will cause the periodic jumping.The guided wave signal can be identified according to the periodic jumping.The increases or decreases of the driving force amplitude with the interference of the guided wave signal heavily depends on the phase between the intercepted signal and the periodic driving force.The phases satisfying the conditions of the increase and decrease of the driving force are out of phase with each other,whose difference is approximately π in the same period.Based on these conclusion,two detection models of small-scale periodic state and large-scale periodic state are constructed,and the effectiveness of the models is checked for identifying the guided wave signal by numerical and experimental studies.At the same time,the anti-noise interference ability of the two models are compared.The results show that the small-scale periodic state detection model has the advantages of unique results and strong anti-noise ability.This method can effectively improve the sensitivity of ultrasonic guided wave detection of small defects.