The wind turbulence inflows specified in current wind turbine design standards and turbine response simulation tools are modeled as stationary Gaussian processes. Field measurement data, however, suggest that wind turbulence in complex terrain exhibits non- Gaussian characteristics. This study presents a comprehensive investigation on extreme response of operational wind turbines to non-Gaussian wind field. The non-Gaussian wind field with specified non-Gaussian statistics and power spectral characteristics are generated using translation process theory. The wind turbine response time histories at each wind speed bin from cut-in to cut-out wind speed are simulated. The extreme response distributions conditional on wind speed are determined from the simulation data using global maxima method and process model method. The overall extreme response distribution is then calculated by further integrating the distribution of mean wind speed, which is used to quantify the extreme responses with various mean recurrence intervals (MRIs). The predicted extreme responses under non-Gaussian wind inflow are compared to those under Gaussian wind inflow. The results showed that the non- Gaussian characteristics of wind inflows can result in considerably larger extreme responses of wind turbines.