This paper presents a recursive least squares algorithm for the fractional order Hammerstein-Wiener system. Considering the coupled terms of the model, the hierarchical identification principle is employed to simultaneously identify the coupled parameters. For the unknown intermediate terms existing in the identification model, the idea of auxiliary model identification is harnessed to substitute the intermediate variables with the output generated by an auxiliary model. Subsequently, the hierarchical recursive least squares algorithm on the basis of the auxiliary model is proposed. The simulation results demonstrate the effectiveness of the presented algorithm.