Solar shortwave radiation penetrating the bottom of the ocean mixed layer (Q_pen) is critical for the heat distribution in the upper ocean and the variability of sea surface temperature. Current climate models typically parameterize Q_pen using chlorophyll-a concentration (Chla). However, this approach has an inherent limitation, as Chla cannot accurately account for the contributions of other optically active constituents (such as colored dissolved organic matter (CDOM) and non-algal particles) to underwater light attenuation, which may lead to substantial biases in global ocean simulations. In this study, we constructed a high-quality matchup dataset (N = 218,470) based on MODIS-Aqua and BGC-Argo observations from 2012 to 2024, and systematically evaluated the performance of several classical Q_pen parameterization schemes, including the L05^[1] scheme based on the inherent optical properties (IOPs) of seawater, the single-exponential Chla-based schemes M02^[2] and T23^[3], and the double-exponential Chla-based scheme O03^[4]. The validation results show that the L05 and O03 schemes provide the lowest uncertainties in estimating Q_pen, with median percentage differences (MPD) of 12.3% and -12.6%, respectively. In contrast, the two single-exponential Chla-based schemes (M02 and T23) systematically overestimate Q_pen by 122.0% and 403.3%, respectively. Furthermore, analyses based on a 20-year time series of Q_pen (2002-2023) derived from MODIS-Aqua using the L05 scheme reveal a significant decreasing trend in the global-ocean mean Q_pen (-0.037 ± 0.001 W/m²/yr, p < 0.01). This finding indicates a slight but statistically significant decline in the solar radiation penetrating below the mixed layer over the past two decades, whose implications for upper-ocean ecosystems and thermodynamic regulation warrant further investigation.