Currently, offshore and coastal wind power resources are growing rapidly around the world, especially in China. However, systematic research on the hazard assessment of wind farms under tropical cyclone conditions remains lacking. This study simulated the wind field of tropical cyclones based on a parameterized tropical cyclone wind field model, and analyzed the characteristics of historical tropical cyclones in Donghaitang wind farm (Zhejiang, China). Four extreme tropical cyclone hazard factors including the maximum wind speed (Vmax), maximum duration of wind direction change (Tmax), maximum cumulative wind direction change (Δθmax) and maximum rate of change in wind direction (ΔDmax) were proposed and examined. Then a comprehensive hazard assessment model for wind farms based on the analytic hierarchy process was established, and the risk to the Donghaitang wind farm represented by tropical cyclones during 1949–2021 was evaluated. Results showed that the number and intensity of tropical cyclones made landfall near the coast of Donghaitang wind farm gradually increased with time, which results in a gradual increase in the composite tropical cyclone risk level of the Donghaitang wind farm with time. The numbers and risk levels of tropical cyclones traveling northwestward were much larger than those traveling northward or northeastward. Moreover, the average composite risk index for tropical cyclones passing to the left of the wind farm was 14.3% higher than that for tropical cyclones passing to the right. The large values of Vmax and ΔDmax are main reasons for the high risk of the wind farm, while the other two hazard factors (Tmax, Δθmax) proposed to account for the wind turbine backup power are also of great importance in the design, selection and operation stages of offshore wind turbines. The findings of this study could provide support for hazard assessment of offshore and coastal wind farms exposed to tropical cyclones, including macro site selection of wind farms and type selection of wind turbines.