A reliability assessment method of fatigue life based on the long-term monitoring data is developed for welded details in steel box girder,and the application research is presented with examples of welded rib-to-deck details in Runyang Bridges. Firstly the fatigue damage limit-state function is established based on S-N curves and Miner’s rule,and the probability distribution characteristics of the coefficients in the function are discussed in detail. The uncertainties in fatigue loading effects are mainly studied based on long-term monitoring data. In the traditional studies,only the uncertainty of equivalent stress range is considered in fatigue reliability assessment. However,stress cycle number is also treated as a random variable in this paper because we know traffic flow every day differs in a thousand ways. Then the optimization method is employed to calculate the fatigue reliability. After studying the changing law of the reliability indices with time and the effect of the randomness of stress cycle number on reliability,the effect of the traffic growth on the reliability is studied. This study shows that the uncertainty in the fatigue life of the welded details can be well studied based on structural health monitoring,so it is necessary to carry out long-term strain monitoring of the welded details for accurate fatigue reliability assessment during the whole service period. The reliability of the fatigue life based on the long-term monitoring data is developed for welded details in steel box girder, and the application research is presented with examples of welded rib-to-deck details in Runyang Bridges. -state function is established based on SN curves and Miner’s rule, and the probability distribution characteristics of the coefficients in the function are discussed in detail. The uncertainties in fatigue loading effects are mainly studied based on long-term monitoring data. In the traditional studies , only the uncertainty of equivalent stress range is considered in fatigue reliability assessment. However, stress cycle number is also treated as a random variable in this paper because we know traffic flow every day differs in a thousand ways. Then the optimization method is employed to calculate the fatigue reliability. After studying the changing law of the reliability indices with time and the effect of the the rando mness of stress cycle number on reliability, the effect of the traffic growth on the reliability is studied. This study shows that uncertainty in the fatigue life of the welded is detailed. out long-term strain monitoring of the welded details for accurate fatigue reliability assessment during the whole service period.