Four single-sided natural ventilation experiments were carried out in a test cell, a full-scale outdoor facility in Athens, Greece. During the experiments air velocity data at various heights at the vertical centerline of the opening were collected. Various mathematical approaches were used for the prediction of the air velocity at the door level, using experimentally derived parameters as inputs. Deterministic as well as intelligent techniques were applied in an attempt to determine the potential of each of these approaches in predicting the air velocity with accuracy. In the frame of the first category, correlation techniques, pressure models and CFD models were studied. The intelligent techniques involved an application of fuzzy theory for the prediction of the air velocity at the opening level. The response of each of the above methodologies is discussed in this paper.