The surface roughness is one of the main parts of hand prediction. Classical method of surface roughness measurements is based on the surface profile measurement. Characteristic of roughness is then variation coefficient of surface profile (surface height variation). The main aim of this work is to estimate the surface profile complexity by using variogram (structure function). The surface profile variation is classified to the group according to short‐ and long‐range dependence. The concept of fractal dimension is proposed especially for long‐term correlation cases. The applicability of the proposed approach is demonstrated on the typical heat protective clothing fabrics and compared with the results of surface roughness evaluated by the KES system.