The postulated fatigue crack growth curve (P-curve), which represents the relationship between crack size and fatigue damage, has been developed for Type 316 stainless steel. In this study, further improvement of the P-curve was made. Particularly, the incubation period before initiation of a 0.1 mm depth crack was newly included in the P-curve. Then, the P-curve was extended for thermal fatigue loading, which is the main cause of fatigue damage in nuclear power plant components. The stress and strain gradient in the depth direction derived from thermal convection analysis was reasonably considered in the crack growth prediction. It was shown that the fatigue damage (DF) of a component could be quantified using the P-curve and crack size detected by inspection, or even if no crack is detected, possible DF could be estimated from the detectable crack size of the inspection technique. By incorporating the probability of detection (POD) of an inspection technique into the P-curve, the PD-curve, which is the relationship between the DF and POD, was developed. It was discussed that the PD-curve can be used to show the specifications of inspection techniques necessary for ensuring component integrity.