Abstract: To evaluate the high-/low-cycle fatigue life of dissimilar-steel welded joints and achieve a unified characterization, this study investigates welded joints between No. 35 forged steel/No. 20 cast steel and Q550E high-strength steel. High-cycle and low-cycle fatigue tests were conducted, and finite element analyses were performed to extract the stress response at the weld toe. The objective is to clarify the differences in fracture locations and their transfer behavior under HCF/LCF conditions for different base-metal combinations, and to compare the predictive performance of various equivalent-parameter methods to verify the unified assessment capability of the equivalent pseudo-structural stress method. The results show that, under high-cycle fatigue, fractures occur at the weld toe for both joints. Under low-cycle fatigue, the 35/Q550E joint still fractures at the weld toe, whereas the fracture location of the 20/Q550E joint shifts to the base metal. Using the equivalent pseudo-structural stress method for fatigue-life assessment, 90.9% of the experimental data fall within the scatter band corresponding to a 99% survival probability, and the prediction accuracy and scatter control are significantly better than those of the equivalent structural stress method and the equivalent structural strain method. The study confirms that the proposed method can reliably provide a unified evaluation of the HCF/LCF fatigue life of dissimilar-steel welded joints, offering important support for engineering fatigue design, life prediction, and safety assessment.