Abstract: Polarized light and its imaging technologies are widely used in underwater detection, offering unique advantages for clear imaging in scattering water environments. Due to absorption and particle scattering in water, natural light is weak in underwater conditions, requiring active illumination to achieve high-quality imaging. The divergence angle of the light source significantly affects the intensity and polarization distribution of the illumination field, thereby influencing imaging quality. Traditional polarized Monte Carlo models often overlook the characteristics of light source divergence angles and introduce errors in recording polarization properties. To address this, this study introduces comprehensive improvements to the traditional polarized Monte Carlo model, including the construction of the reference meridional plane, photon initialization methods, photon tracking calculations, and photon detection processes. Using the improved model, the effects of light source divergence angle and transmission distance on the underwater transmission characteristics of polarized light were investigated under uniform and Gaussian illumination conditions. Further simulations analyzed the impact of divergence angles and transmission distances on different imaging methods and image processing algorithms, followed by experimental validation in underwater imaging. The results show that reducing the divergence angle of active illumination light sources enhances the contrast between imaging targets and background stray light by leveraging differences in polarization characteristics, improves imaging contrast, and suppresses noise interference. The proposed full-chain polarized Monte Carlo model and numerical studies provide valuable guidance and theoretical support for the design of active light sources in underwater imaging detection.