CFD computations of particle flows in a human nasal cavity were conducted, using two types of inlet velocity profiles at the nostril openings respectively. One type is the widely used uniform inlet velocity profile while the other is a realistic inlet velocity profile extracted from the external environment that is influenced by facial features. The latter type of inlet velocity profiles were obtained through CFD computations of indoor gas-particle flows around a human head geometry containing detailed facial features in an enclosed space. Comparisons were made between the predicted results of airflow field, particle tracks and particle deposition efficiency in the nasal cavity. It was revealed that although the effects of facial features exist only in a small region (10–20 mm) in front of the face, they lead to complicated, and non-uniform velocity profiles at the nostril openings. This discrepancy leads to different predictions of airflow fields and local particle deposition efficiency in the nasal cavity, especially in the anterior regions such as the vestibule and nasal valve. The results may further lead to discrepant health-risk assessments associated with particle inhalation. Therefore, for CFD simulations of particle deposition in the nasal cavity, the inlet velocity profiles induced by facial features are important for more realistic results.