Wind tunnel testing is the most important and effective approach to study distribution patterns of wind pressure on building surfaces. In this study, we used string in a double closed silent backflow NH-2 type wind tunnel to investigate wind pressure patterns on plastic greenhouse surfaces. The wind tunnel used is 6 m long, 3 m wide and 2.5 m high. Wind speed can be continually adjusted, and maximum wind speed is 90 m/s. Nonuniformity of the flow field at the experimental site was less than 2%, turbulence intensity less than 0.14%, and average air drift angle less than 0.5. The geometrical reduced scale ratio of the experimental model is 1∶6. Ceiling height, shoulder height, width, and length of this model is 0.475 m, 0.25 m, 10 m, and 1.155 m, respectively. The cambered surface of the ceiling is a semi-ellipse, with semi-major axis 0.5 m. There are 192 points on the model surface, of which 63 points are distributed in three rows on the front surface, and 129 points in another three rows on the ceiling surface. The three rows on the front surface are defined as A1, A2, A3. In each row, 21 points are arranged in ascending order, and the interval between two points is 47.5 m. In the A3 row, the distance between each point and the cambered ceiling surface is 10 mm. On the ceiling surface, models were fixed on the turntable of the wind tunnel, and 13 wind direction angles with an interval of 15° were set in sequence from 0° to 180°. We measured surface wind pressure of a plastic greenhouse, analyzed its patterns, then deduced the critical wind speed of wind-related disasters for such greenhouses. Results show that surface wind pressure changed with wind direction angle. At 45° wind angle, edge wind pressure was negative and reached a maximum value when the windward surface met the leeward surface. The windward surface was controlled by wind pressure. On its windward edge, isolines were dense and the wind gradient large. In contrast, the leeward surface was controlled by wind suction; the wind therefore did not change much. For different wind directions, there were areas with zero pressure in windward zones, whereas pressure coefficients were negative in leeward zones. At 45° leeward angle, negative pressure at the edge of the windward roof reached a maximum. We suggest that wind suction is the main influence on both sides of the roof in plastic greenhouses. Minimum critical wind speed was 14.5 m/s. Our study provides helpful scientific advice for protecting plastic greenhouses.