<p>Porous materials find extensive applications in aerospace, biomedical, and mechanical manufacturing fields, and their permeability is governed by the complex interplay of various pore characteristics. This investigation targets stochastically porous AlSi10Mg architectures, i.e. a class of geometrically undefined porous materials whose pore morphology emerges spontaneously from laser powder bed fusion process dynamics rather than predefined geometric configurations. Through systematic characterization and analysis, we elucidate the fundamental structure-permeability relationships governing stochastic metallic porous architectures. This study employed response surface methodology to design and conduct laser powder bed fusion (LPBF) experiments, systematically investigating the effects of laser power, laser scanning speed, and hatching distance on porosity and morphological characteristics. The pore network architecture was quantitatively characterized using micro-computed tomography three-dimensional reconstruction technology. Moreover, the relationship between pore characteristics and permeability was revealed by a fluid seepage test combined with the modified Kozeny-Carman model. The results show that the low energy density process (7.06–16.48&#xa0;J/mm³) can achieve 14.65–52.55% wide-area porosity control, and the obtained pore shape (low sphericity) is mainly irregular. Permeability is positively correlated with porosity, but it is significantly affected by pore characteristics such as sphericity and tortuosity. The error between the permeability prediction value of the modified model and the permeability value of the seepage test is below 8%. This study established a quantitative relationship between the process, pore characteristics, and permeability of porous materials prepared by laser powder bed fusion, which provided a theoretical foundation for controllable manufacturing of high-performance porous aluminum alloy components and their implementation in fluid transport systems.</p>