ESDU 97024 introduces two Fortran programs@ ESDUpacs A9724 and W9724@ for the consistent derivation of an S-N curve from a set of fatigue data. The method on which the programs are based is applicable to data for both metallic and non-metallic materials as well as to data from tests on coupons or specimens representing structural features. The method also takes account of censored data@ for example@ run-outs@ in a rigorous manner. S may be in terms of a maximum@ an alternating or an equivalent stress@ strain or loading with the life@ N@ expressed in terms of number of cycles@ time or some other convenient unit (for example@ repeated program blocks). The cycles may be of either constant or variable amplitude. The derived curve is the median curve@ that is@ it represents a fifty per cent probability of failure. The Fortran programs perform a non-linear constrained optimisation procedure using a maximum likelihood criterion to derive a curve of the form S = P 1 N q + P 2 N r + P 3 . The values chosen for the unknown parameters in the derived curve are those that maximise the probability of the observed data being obtained. A facility is also provided for the user to specify the values of any of the non-zero parameters that can be estimated from experience. An additional option@ which is only available when censored data are involved@ is that of specifying the value of the coefficient of variation of S@ if it is known. The equation gives a convenient means of estimating S from N@ but not vice versa. The program will@ however@ if required@ calculate values of N corresponding to values of S specified by the user. Details of the input data required by the programs are provided@ along with example input and output files. ESDUpac A9724@ which gives only textual output@ is provided in the form of Fortran source code and corresponding executable file. ESDUpac W9724@ which includes a user-friendly interface and gives graphical as well as textual output@ is provided as an executable file along with comprehensive@ context-sensitive@ help.