Previous studies have found that the fatigue life distributions of different metallic materials tend to have different types of distributions with changes in load levels, while the fatigue strengths corresponding to different lifetimes generally follow the same distribution. Based on the concept of equivalent fatigue strength, the data could be combined under different fatigue lifetimes, and the parameter estimation was carried out by the great likelihood method under the assumption that the fatigue strength of overall samples followed normal distribution. Some classical models, such as the Strohmeyer model, often exhibited poor fitting in the transitional segments from medium to high cycles or from medium to low cycles when using the above method. To address this issue, a middlehigh cycle S-N curve with two transition segments was constructed, and a P-S-N curve suitable for engineering design was provided. Fatigue tests incorporating two transition zones were conducted on ZL101A aluminum alloy and BSM590 structural steel, and comparisons with Strohmeyer model revealed that the newly proposed dual transition S-N curve provided the better fitting goodness with medium and small sample sizes. Compared to the fatigue limit calculated using the lifting method, the deviations of fatigue limits derived from the fatigue strength distribution fitting were all within 3%, while the calculation deviations of the newly proposed S-N curve were all less than 1%. In the transitional segment from medium to high cycles, compared with the results calculated using the Dixonmood method, the accuracy of the proposed S-N curve reached 92.2% and 99.9%, respectively.