Nowadays, a lot of bioimpedance devices are proposed and signal processing for them is still under intensive development. In this study, we improved the complex non-linear least squares (CNLS) protocol for non-stationary  bioimpedance data analysis. To be specific, we tested different Monte-Carlo applications for choosing starting point for CNLS approximation – an essential step for every non-linear problem. As a result, we have proved that symbiosis  approach with usage of the CNLS-solution of the previous spectra as starting point and global Monte-Carlo searching of the starting point is the promising powerful combination for such a task.