This paper introduces a new frequency domain technique for modal identification of output-only systems, where modal parameters are estimated without knowing the input excitation. The technique, called Frequency Domain Decomposition (FDD), is an extension of the classical Peak Picking approach but uses Singular Value Decomposition (SVD) to decompose the spectral density function matrix into a set of single degree of freedom (SDOF) systems. This decomposition allows for high-accuracy identification of close modes, even in the presence of strong noise, and clearly indicates harmonic components in the response signals. The method is user-friendly, faster, and simpler compared to other classical and modern identification techniques. The paper demonstrates the effectiveness of FDD through a simulation example of a two-storey building model, showing high accuracy in estimating natural frequencies and damping ratios, and robustness to noise. Additionally, the technique successfully identifies harmonics in the response signals, providing a clear separation between structural and harmonic peaks.This paper introduces a new frequency domain technique for modal identification of output-only systems, where modal parameters are estimated without knowing the input excitation. The technique, called Frequency Domain Decomposition (FDD), is an extension of the classical Peak Picking approach but uses Singular Value Decomposition (SVD) to decompose the spectral density function matrix into a set of single degree of freedom (SDOF) systems. This decomposition allows for high-accuracy identification of close modes, even in the presence of strong noise, and clearly indicates harmonic components in the response signals. The method is user-friendly, faster, and simpler compared to other classical and modern identification techniques. The paper demonstrates the effectiveness of FDD through a simulation example of a two-storey building model, showing high accuracy in estimating natural frequencies and damping ratios, and robustness to noise. Additionally, the technique successfully identifies harmonics in the response signals, providing a clear separation between structural and harmonic peaks.