Based on the improved damage algorithm and multi-lane refined traffic flow simulation, a new method for predicting the fatigue life of concrete bridges was proposed, where the damage under each cyclic loading was introduced into the S-N curve in the improved damage algorithm. The S-N curve of materials under fatigue was modified in order that the predicted fatigue life of materials was closer to the real situation. The Markov chain Monte-Carlo simulation method (MCMC) was used to generate multi-lane fine traffic flow considering the correlation between adjacent models and traffic lanes. First, the accuracy of the improved damage algorithm was verified by the multistage variational fatigue tests of a group of reinforced concrete beams and a group of prestressed concrete beams. Then, the refined process of multi-lane random traffic flow simulation was introduced, and the fatigue life prediction process of concrete bridge based on improved damage algorithm and multi-lane stochastic traffic flow simulation was proposed. Finally, the traffic flow data measured on a highway and a simple supported T beam bridge with a span of 20m were used for example analysis. The results show that the prediction error of the five groups of specimens is significantly lower than that of the conventional damage algorithm. Except that the prediction error of two prestressed concrete beams is bigger (53%~56%), the prediction error of the other three groups of specimens is smaller (less than 8%), indicating that the improved damage algorithm can be used to predict the fatigue life of concrete bridges. In case of analysis, the stress spectrums appear the characteristic of multi-peak distribution, which is similar to the vehicle load distribution, illustrating the rationality of the simulation. According to the improved damage algorithm, when the average annual traffic rate (AAGR) is 0, the fatigue life of the bridge is 77.50 years, which didn't meet the requirements of design service life. When AAGR is 3%, the fatigue life of the bridge is 52.49 years, which decreased by 32.27% compared with that when AAGR is 0.