This study examined the effects of roof shape, specifically roof pitch and eaves, on cross-ventilation performance through wind tunnel experiments that measure static pressure on walls and the detailed wind velocity distribution for isolated building models with basic roof geometries. A flat roof, a gable roof, and two types of shed roof with different orientations were compared. The buildings with a flat or shallow roof pitch (15°) showed no significant change in the pressure difference between windward and leeward facades. However, a steeper roof pitch (30°) caused a significant increase in pressure difference, especially for the shed roof that ascends in the wind direction (shed-A roof). This was caused by larger projected building areas creating stronger recirculation zones behind the building, leading to lower pressure on the leeward facade. The shed-A roof with a pitch of 30° had the best ventilation performance among the tested cases. Eaves generally reduced pressure differences, except for the shed-A roof. The effect of eaves was weaker than reported in previous studies. The experimental results obtained in this study can serve as a validation database for a more detailed study of this problem using computational fluid dynamics.