In this paper, we formulate a new mechanical accelerated life-testing model called Mecha-Statistical model. This model, considered as parametric, is constructed on the basis of an appropriate mechanical damaging model. Both temperature and stress are introduced in the model. After the data sample of a specified class of material has been collected, the frequency distribution of the Number of cycles to failure can be produced. The mean and standard deviation can be estimated. Assuming the lognormal distribution, the relationship between average number of cycles to failure and material characteristics is modeled. The parameters of different accelerated testing models were determined by numerical methods and used to make comparison between the proposed model and the conventional models. The results obtained reveal that the proposed model describes the lifetime testing data of fatigue tests better than the other models, because it depends on: the product, the test method and the accelerating stress.
