The study of passenger spatial behavior patterns in the terminal building is essential for the terminal managers to meet the needs of passengers with different attributes and improve service levels. The research adopts the K-means clustering method to subdivide spatial selection behaviors of domestic departing passengers of Nanjing Lukou International Airport Terminal 2. In the process of segmenting passengers, it clusters behav- ioral paths of domestic departing passengers into five types of spatial behavior patterns and the different distribu- tions of the basic attributes such as gender, age and income of passengers in these five types are analyzed. The research also shows that commercial facilities in terminals are more attractive to female passengers; young and middle-aged passengers are prone to stay in non-functional areas and engage in consumption activities; passen- gers who have not taken flights before have strong interest in various commercial facilities in terminals; and there is a positive correlation between income levels and passengers' choice probabilities of non-functional areas of terminals. Therefore, study on the differences in air passenger spatial behavior patterns has a practical value to improve the management level of Nanjing Lukou International Airport and optimize spatial layout.