The configuration of a building complex can significantly influence the air-infiltration rate, which subsequently affects the energy consumption for heating and cooling. Given the diversity of parameters associated with a building complex, it is challenging to study all combinations. This study proposed the k-means clustering method to identify representative building clusters, thereby facilitating the derivation of correction coefficient for the air-infiltration rate within a complex. Initially, data pertaining to district characteristics, such as the complex density, height, and building aspect ratio, were collected and clustered to obtain several representative groups of building clusters. Subsequently, a validated computational fluid dynamics program with the renormalization group (RNG) k–ε turbulence model was employed to investigate the spatial characteristics of the pressure coefficient (Cp) on building surfaces across various groups within the complex. Finally, the shelter coefficient (SAIR), which represents the ratio of an individual building to those in different groups, was determined to correct for the air-infiltration rate in diverse district types. The results show that SAIR varies considerably (from 0.15 to 0.53) owing to the different densities, heights, and layouts between buildings. The applicability of SAIR was further demonstrated and verified in a case study of a building in Dalian, China. The research results can provide a method for predicting the air-infiltration rate of building complex.