In order to track performance degradation condition of shore bridge key components accurately，an online degradation feature extraction method based on improved symbol sequence entropy and sliding window singular value is proposed. In view of the special character of the shore bridge and the frequent random vibration impact，a threshold factor is introduced and the sequence symbolization method is improved. The information of the amplitude change of the“coarse grain”is also taken into account while retaining the direction information，thereby，improving the ability of filtering impact and characterizing signal complexity. The improved symbol sequence entropy of the vibration signal is calculated online，combining with the sliding window method，the singular value of the improved sequence entropy sequence is extracted cyclically，which is used as the online degradation feature of the key components of the shore bridge. The full-life vibration signals from the lifting gearbox and track hinge point of Shanghai Port Container Terminal are used to analyze the results. The results show that the sliding window singular value can describe the performance degradation degree of the lifting gearbox and track hinge point. The deeper the degradation degree is，the larger the singular value is. It has the advantages of stable and fast calculation. It can also lay a method foundation for accurately tracking and evaluating health status.