Reliability analysis is a crucial phase in assessing the safety status of new and existing structures. One of its applications is to predict the fatigue life of fatigue prone details. Two models are used to formulate the fatigue limit state: S-N curves in combination with Palmgren-Miner damage accumulation rule and linear elastic fracture mechanics using fatigue crack growth rate curves. Within each model, choices must be made on the values of the variables and these choices are sometimes different in different standards. This study investigates the consistency between the standards by determining the failure probability of the different models and values for a transverse butt weld joint under Variable Amplitude Loading. Partial factors required for the design are then derived as a function of the required reliability for each model and associated values. The influence of the uncertainties related to each involved variable is evaluated by performing a sensitivity analysis.