Abstract: In wind climate analysis, historical wind speed data generally comes from archived measurements of meteorological stations. The average time intervals adopted by different meteorological stations are not consistent. The difference in the average time intervals leads to a conversion problem of the average time intervals in the extreme wind speed prediction process when the samples are based on the historical data of wind speed. Continuous observation of wind speed recorded by 11 meteorological stations in normal wind climate areas in the United States was processed to obtain the wind speed sequences at different average time intervals. The extreme wind speeds of return periods were obtained by using the improved independent storm method and the extreme I distribution model. Subsequently, the influence of average time intervals on the prediction results of extreme wind speed was analyzed. The result shows that in normal wind climate areas, the extreme wind speed decreases with the increase in the average time interval, and it approximates an exponential attenuation; the influence of the average time intervals on the extreme wind speed is basically not affected by the return periods; and the extreme wind speed with a sampling average time interval of one minute is about 19% higher than that with a sampling average time interval of ten minutes. Based on these results, the conversion formula of extreme wind speed with different average time intervals in normal wind climate areas was proposed.