Various methods for estimating fatigue life for machine or structural components are compared and discussed. Guidelines are offered to aid in choosing methods or combinations of methods for particular situations, such as different stages of design. The methods discussed include those based on local strain, nominal stress, component test data, and fracture mechanics. It is noted that the exact choice of a method or variation of a method may be an unimportant detail when one considers the uncertainties which often exist as to the actual load history in service or the time dependent effects of a hostile environment. Also, it is important to make maximum use of any component test data or service experience data which are available.