The increasing use of high-power density inverter-fed electrical machines has heightened concerns regarding insulation reliability due to elevated temperature fluctuations and high dv/dt voltages. This paper presents a matrix-based broadband model for stator winding insulation in inverter-fed electrical machines, aimed at improving insulation condition monitoring. A novel matrix-based broadband common-mode model is proposed in this paper which extends previous work by enabling a mathematical analysis of stator winding transfer functions. The model is developed by segmenting the winding into discrete conductor units, incorporating mutual inductance, resistance, and parasitic capacitance effects. High-frequency variations in resistance and inductance due to eddy currents are accounted for using finite element analysis, while a nodal circuit matrix equation is established for impedance calculations. Experimental validation using a 15 kVA synchronous generator demonstrates strong correlation between the model’s predictions and measured impedance spectra across a wide frequency range. The results confirm that this approach accurately reflects insulation degradation effects and provides a robust theoretical foundation for online insulation monitoring. Compared to conventional methods, the proposed model offers enhanced sensitivity to insulation aging locations and facilitates comprehensive frequency-domain analysis for predictive maintenance strategies.