By constructing simple linear regression model in space, global wind power generation potential and wind power density during 2021-2100 are predicted based on the historical data in 1981-2020 including wind speed, wind power density and annual average surface temperature from 1981 to 2020. From 1981 to 2020, the global wind power generation potential and wind power density show a declining trend, whereas the global annual average surface temperature shows an increasing trend, and there is a certain degree of correlation between the two trends. The results of simple linear regression models show that the mean of yearly total continental power generation potential and the annual average surface temperature in 11 regions are well correlated, negatively in most regions. Under different shared socio-economic pathways (SSPs) in the future, there will be large differences in the global total annual average on-shore wind power generation potential. Under the SSP5-8.5 scenario with the fastest temperature rise, by the end of this century, global total on-shore wind power generation potential will be reduced by 3.3% relative to the level in 2020. By contrast, under the SSP1-1.9 scenario with the highest carbon emission reduction, global total on-shore wind power generation potential is likely returning to the level in 2020 as global annual average surface temperature falls in the second half of this century. Compared with the best scenario SSP1-1.9, the wind power density is decreasing on 61% of global lands under the other SSP scenarios. The faster the temperature is rising in a SSP scenario, the more the wind power density values will decline, and the less the potential of wind power generation will be achieved in the future.