To assess the forecast reliability of the CMA-WSP wind speed forecast product at wind farms in Guizhou, an evaluation was conducted on the 72-hour wind speed forecast product. The study utilized actual wind speed data from 12 benchmark wind turbines at the Wujiangyuan Wind Farm in Guizhou from September 1, 2023, to December 31, 2023. The results indicate that: (1) The wind speed forecast product generally aligns with the trends of actual wind speeds, although the forecasted speeds are overall lower than the measured ones. The [第一作者简介：陶勇，工程师，主要从事专业气象服务研究，E-mail：18798879432@163.com 资助项目：中国华电集团有限公司2023年科技项目与技术标准计划（第一批）“气象赋能光伏电站安全、高效预警预测技术研究及应用”(CHDKJ23-02-40) 通讯作者：陈义义，工程师，航空气象预报保障与技术研究，E-mail：243748677@qq.com]correlation coefficients for the forecasted wind speeds 15 minutes and 1 hour ahead are 0.745 and 0.754, respectively. As the forecast horizon extends, the discrepancy between the forecasted and actual wind speeds increases, and the accuracy gradually decreases. (2) The wind speed correction algorithm performs slightly better in forecasting average wind speeds than in forecasting extreme values. (3) In terms of daily distribution, the forecasted average wind speeds are lower than the measured average wind speeds between 2:00 AM and 5:00 PM, and higher during the other periods; the largest difference between the forecasted and measured average wind speeds occurs from 8:00 AM to 12:00 AM. (4) The accuracy of the 24-hour forecasts for six independent plant areas is generally lower than that of the overall wind farm, while the accuracy of the 48-hour and 72-hour forecasts for the Baicaoping and Zu"anshan plant areas is higher than that of the overall wind farm. Overall, the CMA-WSP wind speed forecast correction product performs well in the wind farms of Guizhou, and further in-depth studies can be conducted using data from a longer time series.