Not all interconnected pores in porous asphalt concrete (PAC) are valid for permeability, the main purpose of this paper is to distinguish the valid and invalid interconnected pores, and on this basis, to evaluate the distribution of various types of pores, and to propose a new evaluation index of pore distribution, so as to optimize the permeability performance of PAC and provide a new analysis method for the design of PAC materials. In this study, all the voids in the PAC mixture were found to be interconnected, but only approximately 82.1% of the interconnected pores were valid for permeability. A highly linear correlation between void content and valid interconnected pore content was proved, indicating that an accurate prediction could be made for valid pore content based on a simple air void content laboratory measurement. The quality and size distribution of internal air voids were investigated, and the Weibull function was proved to fit all types of air void gradation, especially for valid pores. An analysis of the distribution of void areas shows that the valid and invalid pores are mainly distributed in the size ranges of 2.36 mm to 26.5 mm and 1.18 mm to 9.5 mm, respectively. Moreover, the area proportions of all voids and valid pores in different size ranges show normal distribution patterns. The uniformity indices of air voids such as Ratio of Position Deviation (RPD), Coefficient of Variation of Quantity (CVQ) and Uniformity Index (UI) were found unsuitable for evaluating PAC mixtures. Therefore, a new indicator known as position eccentricity ratio (PER) was proposed and proved to accurately reflect the void distribution uniformity. The statistical results of PER indicate that the invalid interconnected pores are distributed near the central area of the specimen and that the valid interconnected pores are distributed slightly closer to the outer margin.