Aluminium alloy welded structures are widely used in lightweight engineering applications due to their high strength-to-weight ratio. Unfortunately, their susceptibility to fatigue failure poses significant challenges for engineers, since traditional fatigue life prediction methods struggle to accurately account for the complexities introduced by in-service loads, which are typically multiaxial and Variable Amplitude (VA). This study addresses this challenge by employing the Peak Stress Method (PSM), an engineering technique designed to exploit the Strain Energy Density criterion to efficiently estimate the fatigue strength of welded joints. Specifically, the PSM is a finite element (FE)-based method that rapidly evaluates the Notch Stress Intensity Factors (NSIFs) at critical locations, such as the weld toe and root, modelled as sharp V-notches with null tip radius. The PSM has previously demonstrated strong capabilities in assessing fatigue strength under Constant Amplitude (CA) loading and, in recent developments, the PSM formulation has been adapted for VA loading by integrating a cumulative damage rule. This research further extends the application of the PSM to multiaxial variable amplitude loading conditions for welded joints in aluminium alloys. The proposed approach is validated against experimental data on non-load carrying (nlc) fillet-welded double transverse or inclined attachments made of 6082-T6 aluminium alloy subjected to variable amplitude in-phase multiaxial local stresses. The results highlighted some apparent discrepancies between the theoretical estimations of the PSM and the experimental fatigue lives of joints with inclined attachments, which must be attributed to an extended crack propagation phase. Accordingly, the present work investigates the extent of the fatigue crack propagation phase under both uniaxial and multiaxial as well as CA and VA loading. More in detail, it incorporates experimental measurements made by using the Direct Current Potential Drop (DCPD) method and numerical simulations of the propagating fatigue cracks. The results provide insight into improving the accuracy of fatigue life predictions, proving the PSM to be a reliable tool for the design of aluminium welded structures subject to in-service loads.