In this paper, a comprehensive approach is developed to integrate computati
onal reliability methods and nondestructive inspection (NDI) for fatigue re
liability evaluation. Uncertainties existing in the performance and outcome
of NDI are considered through numerical statistical measures, and the math
ematical relationships among these measures are derived. A new method is pr
oposed to estimate the probability of detection (POD) and confidence interv
al using these relationships. Also, a new method of using the NDI results t
o update the reliability estimation is presented, which accounts for POD, m
easurement accuracy, false call probability, and the probability that a ver
y large flaw may not be detected. The application of the proposed methods i
s illustrated through an example fatigue problem and inspection techniques
with different POD characteristics are compared.