Abstract: The ability of the most recently released high spatio-temporal resolution dataset, the fifth generation of European Centre for Medium-Range Weather Forecasts atmospheric reanalysis data (ERA5), to reproduce near-surface wind speed climatology and the linear trend is analyzed over China’s mainland during the period of 1979–2018. Results indicate that the 10-m and 100-m wind speeds are highly consistent with in situ station observations in both spatial distribution and annual–seasonal–monthly evolutions. The spatial correlation coefficient of the 10-m wind speed climatology between ERA5 and station observation is 0.66. The near-surface wind speed shows obvious regional characteristics with strong wind in Inner Mongolia, west of Northeast China, north of Xinjiang, and west of Tibet Plateau. Wind speeds in the above regions also exhibit a seasonal difference with peak intensity observed in spring. The strongest (weakest) national-averaged wind speed obtained from the station observation, 10-m, and 100-m ERA5 reanalysis is 2.6, 3.0, and 4.5 (2.0, 2.4, and 3.5) m s⁻¹ in April (August), respectively. The 10-m ERA5 monthly-mean wind speeds are 0.3–0.5 m s⁻¹ stronger than those of the observation and 1.2–1.4 m s⁻¹ weaker than those of the 100-m wind speed. Since 1979, the observed wind speed shows an overall decreasing trend of −0.4 m s⁻¹ (39 a) ^–1. This trend is most obvious in spring −0.5 m s⁻¹ (39 a)^–1, with the biggest drop in the winter of 1979–1992 −0.7 m s⁻¹ (14a) ^–1. Since 2013, the surface wind speed has been increasing. ERA5 reanalysis fails to capture the above decreasing trends in both layers. Moreover, the spatial patterns of the linear trend in four seasons are also different from those of the observations. It is concluded that ERA5 performs well in reproducing the near-surface wind climatology over China but fails to capture the long-term trend.