Service loadings in structures and mechanical components can be modelled as random processes. The durability assessment under such complex loadings is commonly approached in time-domain by using counting methods and damage accumulation rules. An alternative approach could be developed in frequency-domain, where the random loading is characterised by its power spectral density. This book aims to provide an overview on methods for fatigue analysis of random loadings, with particular focus on frequency-domain approach. Classical time-domain load characterisation, counting methods and linear damage rule are first reviewed. Then, frequency-domain spectral methods for analysis of stationary random loadings are discussed, with particular emphasis on Gaussian and non-Gaussian load analysis. Application examples are also developed, with both numerical simulations and experimental load measurements. A general comparison of spectral methods is finally presented. This book should help to shed some light on the frequency-domain fatigue analysis of random loadings and it should be especially useful for researcher working in the field of structural and durability assessment under service loadings