A primary issue in bio-materials science is to design materials with ad-hoc properties, depending on the specific application. Among these properties, friction is recognized as a fundamental aspect characterizing materials for many practical purposes. Recently, new and unexpected frictional properties have been obtained by exploiting hierarchical multiscale structures, inspired by those observed in many biological systems. In order to understand the emergent frictional behaviour of these materials at the macroscale, it is fundamental to investigate their hierarchical structure, spanning across different length scales. In this paper, we introduce a statistical multiscale approach, based on a one-dimensional formulation of the spring-block model, in which friction is modeled at each hierarchical scale through the classical Amontons-Coulomb force with statistical dispersion on the friction coefficients of the microscopic components. By means of numerical simulations, we deduce the global statistical distributions 1 of the elementary structure at micrometric scale, and use them as input distributions for the simulations at the next scale levels. We thus study the influence of microscopic artificial patterning on macroscopic friction coefficients. We show that is it possible to tune the friction properties of a hierarchical surface and provide some insight on the mechanisms involved at different length scales.