The absolute stress of steel members is a key parameter for determining the performance of steel structures. Compared with other nondestructive evaluation methods, ultrasonic methods, which correlate material stress with ultrasonic velocity, have received the greatest amount of research attention. In this study, we investigated the measurement of the absolute stress distribution of steel members using two ultrasonic methods: a longitudinal critically refracted (Lcr) wave method and a shear wave method. The Lcr wave is generated from the longitudinal wave mode conversion and exhibits the greatest sensitivity to stress. The shear waves are generated by the birefringence effect, and their synthesis signal spectrum exhibits a minimum that varies with stress. A comparison of the two absolute stress evaluation methods is performed. Specifically, four steel members with identical dimensions and materials are used to investigate the discreteness of the calibrated parameters. The uniaxial absolute stress distributions of two steel members with variable cross sections are measured using the two methods and verified using the traditional strain gauge method. The results show that the uniaxial stress distributions of the two steel members can be evaluated by both the Lcr wave time-of-flight method and the shear-wave spectrum method, although the latter is more accurate for the measurement of stress distribution. Furthermore, the measurement principles, parametric calibrations, sensitivity, accuracy, and repeatability of the two methods are compared, and their applicability is discussed. KEYWORDS acoustoelastic effect, birefringence effect, Lcr wave TOF method, nondestructive stress measurement, shear-wave spectrum method, steel members, uniaxial absolute stress