This paper introduces an implementation developed by the author for processing results from experimental modal tests that exhibit nonlinear behavior. The method is based on the least squares approach in the complex frequency domain and is implemented in Matlab through several key stages. Initially, transfer functions are constructed using the wavelet transformation of the original time signals, a step that notably mitigates the impact of subharmonics, noise, collisions, and other distortions commonly encountered during testing. Following this, stabilization diagrams are generated, providing insights into natural frequencies, damping ratios, and modal participation factors. The frequency domain is then segmented into regions of constancy near resonance frequencies, where smoothing functions of varying polynomial orders are applied. Finally, a channel-by-channel analysis of damping ratios is performed on the smoothed data, effectively excluding channels that distort the determination of damping ratios. The validation of this implementation involved comparing the processing results of noisy responses from experimental modal tests with those obtained using the Polymax method from Siemens’ LMS Test.Lab software. The validation confirmed that this implementation improves the accuracy of modal parameter determination for noisy signals.