In this paper, surrogate modelling of on-chip spiral inductors along with their inherent physical characteristics is presented. Taking the skin effect into consideration, enhanced single pi model is employed to account for parasitic effects manifesting themselves at higher frequencies. Based on the circuit representation using Y-parameters. both shunt and differential models are explained by means of frequency-dependent impedance and quality factor graphics. In order to extract physical parameters from a predetermined inductance-frequency curve, we offered a novel algorithm which produces distinct combinations of model parameters that evolve through iterations until the result converges to the predetermined data within a 5% error margin. This methodology is a variability analysis approach that can explain which model parameters yield low Qfactors due to process variations during integrated inductor manufacturing.