This thesis presents the evaluation of fatigue lives for six steel overpass bridges which were instrumented by attaching strain ages at the bottom flanges in midspan of the main girders. Measured data were sent respectively through dynamic strain amplifier, low pass filter, A/D converter and stored in computer. All passing vehicles were counted continuously in 72 hours. The strain signals were converted to stress ranges by adopting Rainflow Counting Method, Miner's Law and Hook's Law. Using the obtained stress ranges and traffic conditions the remaining lives of the bridges were evaluated using the method proposed by AASHTO and Palmgren-Miner method. The remaining lives of the overpass bridges, calculated from AASHTO were found to be longer than those from the Palmgren-Miner method. It was also found that the maximum stress, measured in all bridges, may reach about 0.8 Fy as a result of a 67.5 tons full-trailor passing. Finally, it was found that the fatigue truck specified by AASHTO was not appropriate for the traffic condition in Thailand. This because the 6-wheel truck and bus, which are the main source of damage in Bangkok, are not included in life evaluation in the AASHTO specification. This thesis proposes the use of 5 different trucks consisting of 6-wheel truck, bus, 10-wheel truck, semi-trailer and full-trailer in life evaluation. Employing axle distances and weight distributions given by Department of Highway (Thailand), it was shown that the proposed trucks yield more reasonable fatigue life prediction than that calculated from the fatigue truck of AASHTO.