In consideration of the important role that bridges play as transportation infrastructures, their safety, durability and
serviceability have always been deeply concerned. Structural Health Monitoring Systems (SHMS) have been installed to
many long-span bridges to provide bridge engineers with the information needed in making rational decisions for
maintenance. However, SHMS also confronted bridge engineers with the challenge of efficient use of monitoring data.
Thus, methodologies which are robust to random disturbance and sensitive to damage become a subject on which many
researches in structural condition assessment concentrate. In this study, an innovative probabilistic approach for
condition assessment of bridge structures was proposed on the basis of long-term strain monitoring on steel girder of a
cable-stayed bridge. First, the methodology of damage detection in the vicinity of monitoring point using strain-based
indices was investigated. Then, the composition of strain response of bridge under operational loads was analyzed.
Thirdly, the influence of temperature and wind on strains was eliminated and thus strain fluctuation under vehicle loads
is obtained. Finally, damage evolution assessment was carried out based on the statistical characteristics of rain-flow
cycles derived from the strain fluctuation under vehicle loads. The research conducted indicates that the methodology
proposed is qualified for structural condition assessment so far as the following respects are concerned: (a) capability of
revealing structural deterioration; (b) immunity to the influence of environmental variation; (c) adaptability to the
random characteristic exhibited by long-term monitoring data. Further examination of the applicability of the proposed
methodology in aging bridge may provide a more convincing validation.