Although intensive research has been carried out to understand the fatigue behavior of steel notched components, under variable amplitude loading, no definite and general robust models have been derived so far. Therefore, every effort to augment the knowledge in this topic is welcomed. Within this context, existing variable amplitude data, derived by the authors for a notched low carbon pressure vessel steel (P355NL1) flat plate, is used to assess a local approach to fatigue. A linear damage summation framework, supported by elastoplastic finite element analyses, is used. Several variable amplitude loadings were selected and analyzed, using alternative configurations of kinematic hardening plasticity models (e.g., Chaboche’s model with distinct constants superposition). The predictions are assessed using available experimental data and data derived with simplified empirical elastoplastic tools. This paper highlights the difficulties of performing such elastoplastic analysis and compares the obtained results with those obtained using more classical tools for elastoplastic analysis (Glinka and Seeger–Heuler). It was found that fatigue predictions based on an elastoplastic finite element analysis, made using the Chaboche’s model, were significantly more accurate than predictions based on simplified elastoplastic analysis. These results have important practical relevance.