Abstract: The quality factor Q of subsurface media is a critical parameter that characterizes the attenuation of seismic wave energy and plays a pivotal role in reservoir prediction, hydrocarbon resource evaluation, and subsurface structural imaging. Although the spectral-ratio method is a standard approach for Q estimation, its accuracy degrades significantly at low signal-to-noise ratio (SNR) and nonstationary signals, limiting its applicability in reservoir characterization. To address these challenges, we propose a novel Q-estimation method that integrates an improved S-transform (IST) with shaping regularization. First, we obtained a high-precision time-frequency spectrum of the target signal by introducing a nonlinear window function scaling factor into the IST. Subsequently, we reformulated the division operation inherent in the spectral ratio method as a quasi-inversion problem. Based on the principles of shaping regularization, a two-dimensional triangular smoothing operator was constructed to impose continuous prior constraints in both the spatial and frequency domains. Through iterative optimization, random noise and discrete anomalies were effectively suppressed, yielding high-precision and stable Q estimates. Tests on noisy synthetic models demonstrated that the proposed method maintains stability even in low-SNR environments. Application to field seismic data further revealed that the estimated Q values correlated well with the water-saturation curves, successfully identifying thin, tight sandstone reservoirs. Consequently, this method provides a reliable and efficient technical solution for quality factor extraction and reservoir prediction under low-SNR conditions.