We studied the microstructural evolution of 2.25Cr-1Mo steels subjected to tensile creep at 923 K through monitoring of shear-wave attenuation and velocity, using electromagnetic acoustic resonance (EMAR). Contactless transduction based on the magnetostrictive mechanism is the key to establishing a monitor for microstructural change in the bulk of the metals with a high sensitivity. In the short interval, 50 to 60 pct of the creep life, attenuation experiences a peak, being independent of the applied stress. This novel phenomenon is interpreted in terms of the drastic change in dislocation mobility and rearrangement, which is supported by transmission electron microscopy (TEM) observations for dislocation structure. At this particular period, the dense dislocation structure starts to transform to subgrain boundaries, which temporally accompanies long, free dislocation, absorbing much ultrasonic energy to produce the attenuation peak. The EMAR has the potential to assess the damage advance and to predict the remaining creep life of metals.way of evaluating the internal damage state with comparatively simple and easy instrumentation, while many others, including X-ray, replication, hardness, and magnetic properties, merely inspect the surface or near-surface region of the materials.Ultrasonics have been studied for estimating the creep damage. [5][6][7][8][9][10][11] In these applications, the ultrasonic attenuation and velocity have been measured with a contacting piezoelectric transducer. In some cases, the velocity decreased with damage accumulation; in others, this tendency was not observed. The velocity was often more sensitive to the damage than the attenuation. The creep mechanism is dominated by dislocations, which also dissipate acoustic energy causing ultrasonic attenuation (or internal friction). However, the attenuation changes measured using the conventional transducer cannot exactly reflect the creep damage because the as-measured attenuation includes not only the attenuation of the sample, but also the damping through the transducer, the couplant and the buffer, the reflection and transmission losses at the interface, and the energy leakage into the transducer. Because the attenuation within the sample has to be considered in evaluating the creep damage, these extra losses are much greater than the attenuation within the sample [12]