During service, notched designed components such as steam generators in the nuclear power plant usually experience fatigue damage at elevated temperatures, due to the repeated cyclic loadings during start-up and shut-down operations. Under such extreme conditions, the durability of these components is highly-affected. Besides, to assess the fatigue life of these components, a reliable determination of the local stress-strain at the notch-tips is needed. In this work, the maximum strains of circumferentially notched cylindrical specimens were calculated using the most commonly known analytical methods, namely Neuber's rule, modified Neuber's rule, Glinka's rule, and linear rule, with notch root radius of 1.25, 2.5, and 5 mm, made of modified 9Cr–1Mo steel at 550 °C, and subjected to nominal stress amplitudes of ±124.95, ±149.95, and ±174.95 MPa. The calculated local strains were compared to those obtained from Finite Element Analysis (FEA). It was found that all the analytical approximations provided unreliable local strains at the notch-tips, resulting in an overestimation or underestimation of the fatigue life. Therefore, a mathematical model that predicts the fatigue lives for 9Cr–1Mo steel at elevated temperature was proposed in terms of the applied stress amplitude and the fatigue stress concentration factor. The calculated fatigue lifetimes using the proposed model are found to be in good agreement with those obtained experimentally from the literature with relative errors, when the applied stress amplitude is ±149.95 MPa, are of 1.97%,–8.67%, and 13.54%, for notch root radii of 1.25, 2.5, and 5 mm, respectively.