Using oneyear meteorological observation data from the observation towers at the Zuoquan and Wangqu power plants of Shanxi Province and the conventional meteorological data from neighboring weather stations in same time, some technical issues, such as the selection of typical years and contrast stations with good correlation, are discussed. Through analyzing the gradient meteorological data around the aircooling systems, in combination with the longrange meteorological element series of the selected contrast meteorology stations, the wind and temperature fields around the power plants are reconstructed by means of the related statistical methods. Some problems existed in the wind data transformation are discussed. The results indicate that when there is poor correlation between two wind data series, it is better to add and revises the wind data by using conditional probability, linear regression, and wind vector methods. The findings are of certain significance to the reconstruction of historical wind and temperature fields.