Objective: Existing specifications inadequately address the determination of wind vibration coefficients for flexible photovoltaic supports, and research on interference effects in multi-row structures as well as coupled vibrations with multiple degrees of freedom is limited. This study investigates the wind-induced vibration response and fluctuating wind loads on flexible photovoltaic supports through wind tunnel experiments. Furthermore, a method for determining wind vibration coefficients is developed by considering the coupling effects of torsion, vertical, and bending modes, and the results are compared with those of five international codes. The aim is to provide a highly accurate scientific basis for evaluating wind vibration coefficients in practical engineering applications and to support the wind-resistant structural design of highly flexible photovoltaic systems. Methods: An aeroelastic model of a double-row, three-span photovoltaic system was designed, fabricated, and installed in a wind tunnel test environment. Vibration measurement tests were performed under a wide range of inclination angles (from -30° to 30°), attack angles (from -60° to 60°), and flow conditions, including uniform and turbulent flow fields. The tests were designed to systematically examine the interference effects among upstream and downstream photovoltaic rows and the influence of flow characteristics on the wind-induced vibration behavior of the structure. Using vertical and torsional displacements as primary indicators, the wind vibration coefficient was determined by the envelope value method to capture the maximum response range. Results: For nonzero inclination angles, the presence of upstream photovoltaic panels reduced the average displacement response of downstream panels by approximately 30.00% because of shielding and aerodynamic interference. Under turbulent flow conditions, the average displacement trend of the photovoltaic supports was highly complex and irregular, but overall response amplitudes were relatively small. The pulsating displacement amplitudes in the turbulent fields were generally greater than those under uniform flow conditions. Torsional vibration dominated the dynamic response characteristics, while the wind vibration coefficient exhibited a distinctive "decrease-increase-decrease" pattern with varying inclination angles, peaked at attack angles of ±45°, and showed a clear symmetrical distribution. The measured wind vibration coefficients ranged from 1.50-3.00 at different inclination angles, mostly falling between the Japanese and British code values. The values under varying attack angles ranged from 4.50-5.50, which significantly exceeded those calculated according to the existing specifications. Conclusions: The wind vibration coefficient value prescribed by the Chinese standard (1.00) is significantly underestimated and does not reflect the actual wind-induced dynamic behavior of flexible photovoltaic structures. For highly flexible photovoltaic systems, it is strongly recommended that wind tunnel tests or dynamic time-history analyses be conducted to determine the wind vibration coefficient accurately. In addition, adequate safety redundancy must be incorporated into the structural design process. Existing design codes should incorporate structural flexibility correction factors to improve the accuracy and applicability of the prescribed wind vibration coefficient values. In regions frequently affected by typhoons, conservative values from Japanese standards are recommended as a reference. In typical wind zones, an intermediate value between experimental values and those provided by British and American standards is advised for reliable and efficient design.