For reasonable fatigue design and estimation of fatigue durability considered fatigue strength and stiffness of the automotive body structure, many fatigue data must be insured according to the shapes, materials, and welding conditions of the spot-welded lap joints. In general, the most typical and traditional method for fatigue design of the spot-welded thin sheet structure is curve of S-N curve. By the way, these curves are varied according to the welding condition, material, joint geometry, and fatigue loading condition. Thus, in order to determine the fatigue design criterion for spot-welded joint having specific dimension and geometry, additional fatigue tests according to those factors are generally required. This indicates that a lot of cost and time for it have to be consumed.In this paper, an economical method improving such problems was proposed. And, its reliability and performance were verified by the theory of Weibull probability distribution. As a mechanical parameter to determine the fatigue design criterion of spot-welded structure, the maximum stress equation was defined by using artificial neural network. By using the maximum stress equation and the fatigue data previously obtained as the fatigue test results, the fatigue design criterion of spot-welded joint could be predicted without any additional fatigue tests. The results predicted by this method showed good agreement with actual fatigue test results.